Users' Manual

Users' Manual - IRIT

A Solid modeling Program

(C) Copyright 1989, 1990-1996 Gershon Elber

EMail:	gershon@cs.technion.ac.il
Join IRIT mailing list:	gershon@cs.technion.ac.il
Mailing list:	irit-mail@cs.technion.ac.il
Bug reports:	irit-bugs@cs.technion.ac.il
WWW Page:	http://www.cs.technion.ac.il/~irit

This manual is for IRIT Version 7.0

 Copyright (C) 1989, 1990-1997 Gershon Elber

 EMail: gershon@cs.technion.ac.il

 Join IRIT mailing list: gershon@cs.technion.ac.il

 Mailing list: irit-mail@cs.technion.ac.il

 Bug reports: irit-bugs@cs.technion.ac.il

 WWW Page: http://www.cs.technion.ac.il/~irit

Introduction

IRIT is a solid modeler developed for educational purposes. Although small, it is now powerful enough to create quite complex scenes.

IRIT started as a polygonal solid modeler and was originally developed on an IBM PC under MSDOS. Version 2.0 was also ported to X11 and version 3.0 to SGI 4D systems. Version 3.0 also includes quite a few free form curves and surfaces tools. See the UPDATE.NEW file for more detailed update information. In Version 4.0, the display devices were enhanced, freeform curves and surfaces have further support, functions can be defined, and numerous improvement and optimizations are added.

Copyrights

BECAUSE IRIT AND ITS SUPPORTING TOOLS AS DOCUMENTED IN THIS DOCUMENT ARE LICENSED FREE OF CHARGE, I PROVIDE ABSOLUTELY NO WARRANTY, TO THE EXTENT PERMITTED BY APPLICABLE STATE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING, I GERSHON ELBER PROVIDE THE IRIT PROGRAM AND ITS SUPPORTING TOOLS "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THESE PROGRAMS IS WITH YOU. SHOULD THE IRIT PROGRAMS PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, REPAIR OR CORRECTION.

IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW WILL GERSHON ELBER, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY LOST PROFITS, LOST MONIES, OR OTHER SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OR INABILITY TO USE (INCLUDING BUT NOT LIMITED TO LOSS OF DATA OR A FAILURE OF THE PROGRAMS TO OPERATE WITH PROGRAMS NOT DISTRIBUTED BY GERSHON ELBER) THE PROGRAMS, EVEN IF YOU HAVE BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES, OR FOR ANY CLAIM BY ANY OTHER PARTY.

IRIT is a freeware solid modeler. It is not public domain since I hold copyrights on it. However, unless you are to sell or attempt to make money from any part of this code and/or any model you made with this solid modeler, you are free to make anything you want with it.

IRIT can be compiled and executed on numerous Unix systems as well as OS2, Windows NT/95 and AmigaDOS. However, beware the MSDOS support is fading away.

You are not obligated to me or to anyone else in any way by using IRIT. You are encouraged to share any model you made with it, but the models you made with it are yours, and you have no obligation to share them. You can use this program and/or any model created with it for non commercial and non profit purposes only. An acknowledgement on the way the models were created would be nice but is not required.

Command Line Options and Set Up

The IRIT program reads a file called irit.cfg each time it is executed. This file configures the system. It is a regular text file with comments, so you can edit it and properly modify it for your environment. This file is being searched for in the directory specified by the IRIT_PATH environment variable. For example 'setenv IRIT_PATH /u/gershon/irit/bin/'. Note IRIT_PATH must terminate with '/'. If the variable is not set only the current directory is being searched for irit.cfg.

In addition, if it exists, a file by the name of iritinit.irt will be automatically executed before any other '.irt' file. This file may contain any IRIT command. It is the proper place to put your predefined functions and procedures if you have some. This file will be searched much the same way IRIT.CFG is. The name of this initialization file may be changed by setting the StartFile entry in the configuration file. This file is far more important starting at version 4.0, because of the new function and procedure definition that has been added, and which is used to emulate BEEP, VIEW, and INTERACT for example.

The solid modeler can be executed in text mode (see the .cfg and the -t flag below) on virtually any system with a C compiler.

Under all systems the following environment variables must be set and updated:

path	Add to path the directory where IRIT's binaries are.
IRIT_PATH	Directory with config., help and IRIT's binary files.
IRIT_DISPLAY	The graphics driver program/options. Must be in path.
IRIT_BIN_IPC	If set, uses binary Inter Process Communication.

For example,

set path = (path /u/gershon/irit/bin) setenv IRIT_PATH /u/gershon/irit/bin/ setenv IRIT_DISPLAY "xgldrvs -s-" setenv IRIT_BIN_IPC 1

to set /u/gershon/irit/bin as the binary directory and to use the sgi's gl driver. If IRIT_DISPLAY is not set, the server (i.e., the IRIT program) will prompt and wait for you to run a client (i.e., a display driver). if IRIT_PATH is not set, none of the configuration files, nor the help file will be found.

If IRIT_BIN_IPC is not set, text based IPC is used, which is far slower. No real reason not to use IRIT_BIN_IPC, unless it does not work for you.

In addition, the following optional environment variables may be set.

IRIT_MALLOC	If set, apply dynamic memory consistency testing.
	Programs will execute much slower in this mode.
IRIT_MALLOC_PTR	Set to a pointer address and the program will scream
	once this pointer is allocated, if IRIT_MALLOC is set.
IRIT_NO_SIGNALS	If set, no signals are caught by IRIT.
IRIT_SERVER_HOST	Internet Name of IRIT server (used by graphics driver).
IRIT_SERVER_PORT	Used internally to the TCP socket number. Should not
	be set by users.
LD_LIBRARY_PATH	If shared libraries are created, this variable must be
	updated to point to the shared libraries' directory.

For example,

setenv IRIT_MALLOC 1 setenv IRIT_MALLOC_PTR 1234567890 setenv IRIT_NO_SIGNALS 1 setenv IRIT_SERVER_HOST irit.cs.technion.ac.il

IRIT_MALLOC is useful for programmers, or when reporting a memory fatal error occurrence. This variable, when set as a non zero value will activate the following (hexadecimal bit settings with any combination of the following):

0x01	A test for overwriting before the dynamic memory
	allocated or immediately after it. Cheap in time.
0x02	Savings of all allocated objects in a table for the
	detection of freeing unallocated objects and consistency
	of the entire dynamic memory. Expensive in time.
0x04	Zeros every freed object, once it is freed.

IRIT_NO_SIGNALS is also useful for debugging when contorl-C is used within a debugger. The IRIT_SERVER_HOST/PORT controls the server/client (IRIT/Display device) communication.

IRIT_SERVER_HOST and IRIT_SERVER_PORT are used in the unix and Window NT ports of IRIT.

See the section on the graphics drivers for more details.

A session can be logged into a file as set via LogFile in the configuration file. See also the LOGFILE command.

The following command line options are available:

IRIT [-t] [-z] [file.irt]

-t	Puts IRIT into text mode. No graphics will be displayed and
	the display commands will be ignored. Useful when one needs to
	execute an irt file to create data on a tty device...
-z	Prints usage message and current configuration/version
	information.
file.irt	A file to invoke directly instead of waiting to input from
	stdin.

IBM PC OS2 Specific Set Up

Under OS2 the IRIT_DISPLAY environment variable must be set (if set) to os2drvs.exe without any option (-s- will be passed automatically). os2drvs.exe must be in a directory that is in the PATH environment variable. IRIT_BIN_IPC can be used to signal binary IPC which is faster. Here is a complete example:

set IRIT_PATH=c:\irit\bin\ set IRIT_DISPLAY=os2drvs -s- set IRIT_BIN_IPC=1

assuming the directory specified by IRIT_PATH holds the executables of IRIT and is in PATH.

If IRIT_BIN_IPC is not set, text based IPC is used which is far slower. No real reason not to use IRIT_BIN_IPC unless it does not work for you.

The OS2 executables are typically built using the EMX port of gnu C compiler. The distribution of the executables does not include the run time library of EMX and attempt to run IRIT will fail with error message like "File EMX does not exist". You will need to get the run time from

ftp to ftp-os2.nmsu.edu (aliased also as hobbes.NMSU.Edu) cd to os2/unix/emx09c (or a newer version number/level) get emxrt.zip and place its dlls in a place they would be found.

IBM PC Window NT Specific Set Up

The NT port uses sockets and is, in this respect, similar to the unix port. The envirnoment variables IRIT_DISPLAY, IRIT_SERVER_HOST, IRIT_BIN_IPC should all be set in a similar way to the Unix specific setup. As a direct result, the server (IRIT) and the display device may be running on different hosts. For example the server might be running on an NT system while the display device will be running on an SGI4D exploiting the graphic's hardware capabilities. Here is a complete example:

set IRIT_PATH=c:\irit\bin\ set IRIT_DISPLAY=wntgdrvs -s- set IRIT_BIN_IPC=1

Unix Specific Set Up

Under UNIX using X11 (x11drvs driver) add the following options to your .Xdefaults. Most are self explanatory. The Trans attributes control the transformation window, while the View attributes control the view window. SubWin attributes control the subwindows within the Transformation window.

#if COLOR
iritTransBackGround:	NavyBlue
iritTransBorderColor:	Red
iritTransBorderWidth:	3
iritTransTextColor:	Yellow
iritTransSubWin*BackGround:	DarkGreen
iritTransSubWin*BorderColor:	Magenta
iritTransGeometry:	=150x500+500+0
iritTransCursorColor:	Green
iritViewBackGround:	NavyBlue
iritViewBorderColor:	Red
iritViewBorderWidth:	3
iritViewGeometry:	=500x500+0+0
iritViewCursorColor:	Red
irit*MaxColors:	15
#else
iritTransGeometry:	=150x500+500+0
iritTransBackGround:	Black
iritViewGeometry:	=500x500+0+0
iritViewBackGround:	Black
irit*MaxColors:	1
#endif

First Usage

Commands to IRIT are entered using a textual interface, usually from the same window the program was executed from.

Some important commands to begin with are,

1. include("file.irt"); - will execute the commands in file.irt. Note include can be recursive up to 10 levels. To execute the demo (demo.irt) simply type 'include("demo.irt");'. Another way to run the demo is by typing demo(); which is a predefined procedure defined in iritinit.irt.

2. help(""); - will print all available commands and how to get help on them. A file called irit.hlp will be searched as irit.cfg is being searched (see above), to provide the help.

3. exit(); - close everything and exit IRIT.

Most operators are overloaded. This means that you can multiply two scalars (numbers), or two vectors, or even two matrices, with the same multiplication operator (*). To get the on-line help on the operator '*' type 'help("*");'

The best way to learn this program (like any other program...) is by trying it. Print the manual and study each of the commands available. Study the demo programs (*.irt) provided as well.

The "best" mode to use irit is via the emacs editor. With this distribution an emacs mode for irit files (irt postfix) is provided (irit.el). Make your .emacs load this file automatically. Loading file.irt will switch emacs into Irit mode that supports the following three keystrokes:

Meta-E	Executes the current line
Meta-R	Executes the current Region (Between Cursor and Mark)
Meta-S	Executes a single line from input buffer

The first time one of the above keystrokes is hit, emacs will fork an Irit process so that Irit's stdin is controlled via the above commands. This emacs mode was tested under various unix environments, under OS2 2.x/3.x, and under Windows NT 3.x/4.x and Windows 95.

Data Types

These are the Data Types recognized by the solid modeler. They are also used to define the calling sequences of the different functions below:

ConstantType	Scalar real type that cannot be modified.
NumericType	Scalar real type.
VectorType	3D real type vector.
PointType	3D real type point.
CtlPtType	Control point of a freeform curve or surface.
MatrixType	4 by 4 matrix (homogeneous transformation matrix).
PolygonType	Object consists of polygons.
PolylineType	Object consists of polylines.
CurveType	Object consists of curves.
SurfaceType	Object consists of surfaces.
TrimSrfType	Object consists of trimmed surfaces.
TriSrfType	Object consists of triangular surfaces
TrivarType	Object consists of trivariate functions.
GeometricType	One of Polygon/lineType, CurveType, SurfaceType,
	TrimSrfType, TrivarType, TriSrfType.
InstanceType	Object with a GeometryType and a Transformation.
GeometricTreeType	A list of GeometricTypes or GeometricTreeTypes.
StringType	Sequence of chars within double quotes - "A string".
	Current implementation is limited to 80 chars.
AnyType	Any of the above.
ListType	List of (any of the above type) objects. List
	size is dynamically increased, as needed.

Although points and vectors are not the same, IRIT does not destinguish between them, most of the time. This might change in the future.

Commands summary

These are all the commands and operators supported by the IRIT solid modeler:

+	CHDIR	CZEROS	LN	SBEZIER	SYMBSUM
-	CINFLECT	DSTPTLN	LOAD	SBISECTOR	SYSTEM
*	CINTERP	DSTPTPLN	LOFFSET	SBSPLINE	TAN
/	CIRCLE	DSTLNLN	LOG	SCALE	TBEZIER
^	CIRCPOLY	ERROR	LOGFILE	SCRVTR	TBSPLINE
=	CLNTCLOSE	EVOLUTE	MERGEPOLY	SDERIVE	TCRVTR
==	CLNTEXEC	EXIT	MESHSIZE	SDIVIDE	TDERIVE
!=	CLNTREAD	EXP	MOFFSET	SEDITPT	TEDITPT
<	CLNTWRITE	EXTRUDE	MOMENT	SEVAL	TEVAL
>	CMESH	FFCOMPAT	MRCHCUBE	SFOCAL	TEXTGEOM
<=	CMORPH	FFCTLPTS	MSLEEP	SFROMCRVS	TFROMSRFS
>=	CMULTIRES	FFEXTREME	NIL	SGAUSS	TIME
ABS	CNORMAL	FFKNTVEC	NTH	SIN	TINTERP
ACOS	COERCE	FFMATCH	OFFSET	SINTERP	THISOBJ
ADAPISO	COLOR	FFMERGE	ORTHOTOMC	SIZEOF	TMORPH
ARC	COMMENT	FFMSIZE	PAUSE	SMEANSQR	TORUS
AREA	COMPOSE	FFORDER	PCIRCLE	SMERGE	TRAISE
ASIN	CON2	FFPOLES	PDECIMATE	SMORPH	TRANS
ATAN	CONE	FFPTDIST	PDOMAIN	SNOC	TREFINE
ATAN2	CONTOUR	FFPTTYPE	PLN3PTS	SNORMAL	TREGION
ATTRIB	CONVEX	FFSPLIT	POLY	SNRMLSRF	TREPARAM
AOFFSET	COORD	FLOOR	POWER	SPHERE	TRIANGL
AWIDTH	COS	FMOD	PRINTF	SQRT	TRIMSRF
BBOX	COVERISO	FOR	PRISA	SRAISE	TSBEZIER
BOOLONE	COVERPT	FREE	PROCEDURE	SREFINE	TSBSPLINE
BOOLSUM	CPOLY	FUNCTION	PT3BARY	SREGION	TSDERIVE
BOX	CRAISE	GBOX	PTHMSPR	SREPARAM	TSEVAL
BSP2BZR	CREFINE	GETATTR	PTLNPLN	SRINTER	TSNORMAL
BZR2BSP	CREGION	GETLINE	PTPTLN	STANGENT	TVLOAD,
CBEZIER	CREPARAM	GPOLYGON	PTSLNLN	STRIMSRF	VARLIST
CBISECTOR	CROSSEC	GPOLYLINE	RANDOM	STRIVAR	VECTOR
CBSPLINE	CRV2TANS	HELP	RMATTR	SURFPREV	VIEW
CCINTER	CRVLNDST	HERMITE	ROTVEC	SURFREV	VIEWOBJ
CCRVTR	CRVPTDST	HOMOMAT	ROTX	SVISIBLE	VOLUME
CDERIVE	CRVPTTAN	IF	ROTY	SWEEPSRF	WHILE
CDIVIDE	CSURFACE	INCLUDE	ROTZ	SWPSCLSRF
CEDITPT	CTANGENT	INSTANCE	ROTZ2V,	SYMBCPROD
CENVOFF	CTLPT	INTERACT	ROTZ2V2,	SYMBDIFF
CEVAL	CTRIMSRF	IRITSTATE	RULEDSRF	SYMBDPROD
CEXTREMES	CYLIN	LIST	SAVE	SYMBPROD

Functions and Variables

Functions that return a NumericType:

ABS	COS	EXP	POWER	THISOBJ
ACOS	CLNTEXEC	FLOOR	RANDOM	VOLUME
AREA	CPOLY	FMOD	SIN
ASIN	DSTPTLN	LN	SIZEOF
ATAN	DSTPTPLN	LOG	SQRT
ATAN2	DSTLNLN	MESHSIZE	TAN

Functions that return a GeometricType:

ADAPISO	CON2	FFORDER	PTSLNLN	SWEEPSRF
ARC	CONE	FFPOLES	RULEDSRF	SWPSCLSRF
AOFFSET	CONTOUR	FFPTDIST	SBEZIER	SYMBCPROD
BBOX	CONVEX	FFPTTYPE	SBISECTOR	SYMBDIFF
BOOLONE	COORD	FFSPLIT	SBSPLINE	SYMBDPROD
BOOLSUM	COVERISO	GBOX	SCRVTR	SYMBPROD
BOX	COVERPT	GETATTR	SDERIVE	SYMBSUM
BSP2BZR	CRAISE	GETLINE	SDIVIDE	TBEZIER
BZR2BSP	CREFINE	GPOLYGON	SEDITPT	TBSPLINE
CBEZIER	CREGION	GPOLYLINE	SEVAL	TCRVTR
CBISECTOR	CREPARAM	HERMITE	SFOCAL	TDERIVE
CBSPLINE	CROSSEC	INSTANCE	SFROMCRVS	TEDITPT
CCINTER	CRV2TANS	LOFFSET	SGAUSS	TEVAL
CCRVTR	CRVLNDST	MERGEPOLY	SINTERP	TEXTGEOM
CDERIVE	CRVPTDST	MOFFSET	SMEANSQR	TFROMSRFS
CDIVIDE	CRVPTTAN	MOMENT	SMERGE	TINTERP
CEDITPT	CSURFACE	MRCHCUBE	SMORPH	TMORPH
CENVOFF	CTANGENT	NIL	SNORMAL	TORUS
CEVAL	CTRIMSRF	OFFSET	SNRMLSRF	TRAISE
CEXTREMES	CTLPT	ORTHOTOMC	SPHERE	TREFINE
CINFLECT	CYLIN	PCIRCLE	SRAISE	TREGION
CINTERP	CZEROS	PDECIMATE	SREFINE	TREPARAM
CIRCLE	EVOLUTE	PDOMAIN	SREGION	TRIANGL
CIRCPOLY	EXTRUDE	PLN3PTS	SREPARAM	TRIMSRF
CLNTREAD	FFCOMPAT	POLY	SRINTER	TSBEZIER
CMESH	FFCTLPTS	PRISA	STANGENT	TSBSPLINE
CMORPH	FFEXTREME	PROCEDURE	STRIMSRF	TSDERIVE
CMULTIRES	FFKNTVEC	PT3BARY	STRIVAR	TSEVAL
CNORMAL	FFMATCH	PTHMSPR	SURFPREV	TSNORMAL
COERCE	FFMERGE	PTLNPLN	SURFREV	TVLOAD
COMPOSE	FFMSIZE	PTPTLN	SVISIBLE

Functions that create linear transformation matrices:

HOMOMAT	ROTX	ROTZ	ROTZ2V2	TRANS
ROTVEC	ROTY	ROTZ2V	SCALE

Miscellaneous functions:

ATTRIB	ERROR	INCLUDE	NTH	SYSTEM
AWIDTH	EXIT	INTERACT	PAUSE	TIME
CHDIR	FOR	IRITSTATE	PRINTF	VARLIST
CLNTCLOSE	FREE	LIST	PROCEDURE	VECTOR
CLNTWRITE	FUNCTION	LOAD	RMATTR	VIEW
COLOR	HELP	LOGFILE	SAVE	VIEWOBJ
COMMENT	IF	MSLEEP	SNOC	WHILE

Variables that are predefined in the system:

AXES	MACHINE	POLY_APPROX_TOL	VIEW_MAT
DRAWCTLPT	POLY_APPROX_OPT	PRSP_MAT
FLAT4PLY	POLY_APPROX_UV	RESOLUTION

Constants that are predefined in the system:

AMIGA	E3	MAGENTA	PARAM_CENTRIP	SURFACE_TYPE
APOLLO	E4	MATRIX_TYPE	PARAM_CHORD	SUN
BLACK	E5	MSDOS	PARAM_UNIFORM	TRIMSRF_TYPE
BLUE	FALSE	NUMERIC_TYPE	PI	TRIVAR_TYPE
COL	GREEN	OFF	PLANE_TYPE	TRUE
CTLPT_TYPE	HP	ON	POINT_TYPE	UNDEF_TYPE
CURVE_TYPE	IBMOS2	P1	POLY_TYPE	UNIX
CYAN	IBMNT	P2	RED	VECTOR_TYPE
DEPTH	KV_FLOAT	P3	ROW	WHITE
E1	KV_OPEN	P4	SGI	YELLOW
E2	LIST_TYPE	P5	STRING_TYPE

Language description

The front end of the IRIT solid modeler is an infix parser that mimics some of the C language behavior. The infix operators that are supported are plus (+), minus (-), multiply (*), divide (/), and power (^), for numeric operators, with the same precedence as in C.

However, unlike the C language, these operators are overloaded, or different action is taken, based upon the different operands. This means that one can write '1 + 2', in which the plus sign denotes a numeric addition, or one can write 'PolyObj1 + PolyObj2', in which case the plus sign denotes the Boolean operation of a union between two geometric objects. The exact way each operator is overloaded is defined below.

In this environment, reals, integers, and even Booleans, are all represented as real types. Data are automatically promoted as necessary. For example, the constants TRUE and FALSE are defined as 1.0 and 0.0 respectively.

Each expression is terminated by a semicolon. An expression can be as simple as 'a;' which prints the value of variable a, or as complex as:

for ( t = 1.1, 0.1, 1.9, cb1 = csurface( sb, COL, t ): color( cb1, green ): snoc( cb1, cb_all ) );

While an expression is terminated with a semicolon, a colon is used to terminate mini-expressions within an expression.

Once a complete expression is read in (i.e., a semicolon is detected) and parsed correctly (i.e. no syntax errors are found), it is executed. Before each operator or a function is executed, parameter type matching tests are made to make sure the operator can be applied to these operand(s), or that the function gets the correct set of arguments.

The parser is totally case insensitive, so Obj, obj, and OBJ will refer to the same object, while MergePoly, MERGEPOLY, and mergePoly will refer to the same function.

Objects (Variables if you prefer) need not be declared. Simply use them when you need them. Object names may be any alpha-numeric (and underscore) string of at most 30 characters. By assigning to an old object, the old object will be automatically deleted and if necessary its type will be modified on the fly.

Example:

V = sin( 45 * pi / 180.0 ); V = V * vector( 1, 2, 3 ); V = V * rotx( 90 ); V = V * V;

will assign to V a NumericType equal to the sine of 45 degrees, the VectorType ( 1, 2, 3 ) scaled by the sine of 45, rotate that vector around the X axis by 90 degrees, and finally a NumericType which is the dot (inner) product of V with itself.

The parser will read from stdin, unless a file is specified on the command line or an INCLUDE command is executed. In both cases, when the end of file is encountered, the parser will again wait for input from stdin. In order to execute a file and quit in the end of the file, put an EXIT command as the last command in the file.

Operator overloading

Overloading +

The + operator is overloaded above the following domains:

NumericType + NumericType -> NumericType VectorType + VectorType -> VectorType (Vector addition) MatrixType + MatrixType -> MatrixType (Matrix addition) PolygonType + PolygonType -> PolygonType (Boolean UNION operation) CurveType + CurveType -> CurveType (Curve curve profiling) CurveType + CtlPtType -> CurveType (Curve control point profiling) CtlPtType + CtlPtType -> CurveType (Control points profiling) ListType + ListType -> ListType (Append lists operator) StringType + StringType -> StringType (String concat) StringType + RealType -> StringType (String concat, real as int string)

Note: Boolean UNION of two disjoint objects (no common volume) will result with the two objects combined. It is the USER responsibility to make sure that the non intersecting objects are also disjoint - this system only tests for no intersection. Boolean UNION of two polyline objects will merge the list of polylines.

Overloading -

The - operator is overloaded above the following domains:

As a binary operator:

NumericType - NumericType -> NumericType VectorType - VectorType -> VectorType (Vectoric difference) MatrixType - MatrixType -> MatrixType (Matrix difference) PolygonType - PolygonType -> PolygonType (Boolean SUBTRACT operation)

As a unary operator:

- NumericType -> NumericType - VectorType -> VectorType (Scale vector by -1) - MatrixType -> MatrixType (Scale matrix by -1) - PolygonType -> PolygonType (Boolean NEGATION operation) - CurveType -> CurveType (Curve parameterization is reversed) - SurfaceType -> SurfaceType (Surface parameterization is reversed)

Note: Boolean SUBTRACT of two disjoint objects (no common volume) will result with an empty object. For both a curve and a surface parameterization, reverse operation (binary minus) causes the object normal to be flipped as a side effect.

Overloading *

The * operator is overloaded above the following domains:

NumericType * NumericType -> NumericType VectorType * NumericType -> VectorType (Vector scaling) VectorType * CurveType -> CurveType (Inner product projection) VectorType * SurfaceType -> SurfaceType (Inner product projection) VectorType * VectorType -> NumericType (Inner product) MatrixType * NumericType -> MatrixType (Matrix Scaling) MatrixType * PointType -> PointType (Point transformation) MatrixType * CtlPtType -> CtlPtType (Ctl Point transformation) MatrixType * VectorType -> VectorType (Vector transformation) MatrixType * MatrixType -> MatrixType (Matrix multiplication) MatrixType * GeometricType -> GeometricType (Object transformation) MatrixType * ListType -> ListType (Object hierarchy transform.) PolygonType * PolygonType -> PolygonType (Boolean INTERSECTION operation) InstanceType * MatrixType -> InstanceType (Transform of Instance's matrix)

Note: Boolean INTERSECTION of two disjoint objects (no common volume) will result with an empty object. Object hierarchy transform transforms any transformable object (GeometricType) found in the list recursively. Boolean INTERSECTION of two planar (XY plane) polyline objects will compute the intersection points of the two lists of polylines.

Overloading /

The / operator is overloaded above the following domains:

NumericType / NumericType -> NumericType PolygonType / PolygonType -> PolygonType (Boolean CUT operation)

Note: Boolean CUT of two disjoint objects (no common volume) will result with an empty object.

Overloading ^{ }

The ^ operator is overloaded above the following domains:

NumericType ^ NumericType -> NumericType VectorType ^ VectorType -> VectorType (Cross product) MatrixType ^ NumericType -> MatrixType (Matrix to the (int) power) PolygonType ^ PolygonType -> PolygonType (Boolean MERGE operation) StringType ^ StringType -> StringType (String concat) StringType ^ RealType -> StringType (String concat, real as real string)

Note: Boolean MERGE simply merges the two sets of polygons without any intersection tests. Matrix powers must be positive integers or -1, in which case the matrix inverse (if it exists) is computed.

Overloading Equal (Assignments)

Assignments are allowed as side effects, in any place in an expression. If "Expr" is an expression, then "var = Expr" is the exact same expression with the side effect of setting Var to that value. There is no guarantee on the order of evaluation, so using Vars that are set within the same expression is a bad practice. Use parentheses to force the order of evaluation, i.e., "( var = Expr )".

Comparison operators ==, !=, <, >, <=, >=

The conditional comparison operators can be applied to the following domains (o for a comparison operator):

NumericType o NumericType -> NumericType StringType o StringType -> NumericType PointType o PointType -> NumericType VectorType o VectorType -> NumericType PlaneType o PlaneType -> NumericType

The returned NumericType is non-zero if the condition holds, or zero if not. For PointTypes, VectorTypes, and PlaneTypes, only == and != comparisons are valid. This is either the same or different. For NumericTypes and StringTypes (uses strcmp) all comparisons are valid.

Logical operators &&, ||, !

Complex logical expressions can be defined using the logical and (&&), logical or (||) and logical not (!). These operators can be applied to NumericTypes that are considered Boolean results. That is, true for a non-zero value, and false otherwise. The returned NumericType is true if both operands are true for the and operator, at least one is true for the or operator, and the operand is false for the not operator. In all other cases, a false is returned. To make sure Logical expressions are readable, the and and or operators are defined to have the same priority. Use parentheses to disambiguate a logical expression and to make it more readable.

Geometric Boolean Operations

The IRIT solid modeling system supports Boolean operations between polyhedra objects. Freeform objects will be automaticaly converted to a polygonal representation when used in Boolean operations. The +, *, and - are overloaded to denote Boolean union, intersection and subtraction when operating on geometric entities. - can also be used as an unary operator to reverse the object orientation inside out.

IRIT supports Boolean operations on polyhedra models. A polyhedra based model is simply a collection of polygons. While a polyhedra is simply a set of polygons, this set must conform to certain conditions:

Every polygon has known adjacent polygons, for all its edges.

Every polygon has a normal that points into the model. That normal is inside/outside consistent with its adjacent polygons. In other words, for every polygon, one can locally determine the inside or the outside of the model. Moreover, every polygon has neighbors for all its edges, forming a closed object that consistently delineates inside from outside, globally.

The Boolean operations are set operations conducted between two such models, M_1 and M_2, that delineate inside from outside. Boolean Union, Boolean Intersection and Boolean Subtraction are the three common operations that resemble the exact semantic that is expected, when treating M_1 and M_2 as three dimensional point sets.

The Boolean operations can be formulated into a binary tree structure also known as Constructive Solid Geometry (CSG) tree.

Example:

resolution = 20; B = box(vector(-1, -1, -0.25), 2, 1.2, 0.5); C = con2(vector(0, 0, -1.5), vector(0, 0, 3), 0.7, 0.3); D = convex(B - C); E = convex(C - B); F = convex(B + C); G = convex(B * C); tr = rotx( -90 ) * roty( 40 ) * rotx( -30 ); All = list( D * tr * trans( vector( 0.6, 0.5, 0.0 ) ), E * tr * trans( vector( 3.0, 0.0, 0.0 ) ), F * tr * trans( vector( -2.0, 0.0, 0.0 ) ), G * tr * trans( vector( 0.7, -1.0, 0.0 ) ) ) * scale( vector( 0.25, 0.25, 0.25 ) ) * trans( vector( -0.1, -0.3, 0.0 ) ); view_mat = rotx( 0 ); view( list( view_mat, All ), on ); save( "booleans", list( view_mat, All ) );

A complete example to compute the union, intersection and both differences of a box and a truncated cone.

Special cases can be very difficult to handle when considering Boolean operations. Consider an axes parallel bounding cube. Consider a second cube rotated alpha degrees from the first cube. At large angles, the Boolean operations are fairly simple to compute. Nevertheless, as alpha approaches zero, the almost coplanar planes of the two intersecting cubes make it very difficult to robustly and consistently compute their intersection.

There are several flags to control the Boolean operations. See IRITSTATE command for the "InterCrv", "Coplanar", and "PolySort" states.

Priority of operators

The following table lists the priority of the different operators.

Lowest	Operator	Name of operator
priority	,	comma
	:	colon
	&&, \|\|	logical and, logical or
	=,==,!=,<=,>=,<,>	assignment, equal, not equal, less
		equal, greater equal, less, greater
	+, -	plus, minus
	*, /	multiply, divide
Highest	^	power
priority	-, !	unary minus, logical not

Grammar

The grammar of the IRIT parser follows similar guidelines as the C language for simple expressions. However, complex statements differ. See the IF, FOR, FUNCTION, and PROCEDURE below for the usage of these clauses.

Function Description

NumericType returning functions

ABS

NumericType ABS( NumericType Operand )

Returns the absolute value of the given Operand.

ACOS

NumericType ACOS( NumericType Operand )

Returns the arc cosine value (in radians) of the given Operand.

AREA

NumericType AREA( PolygonType Object )

Returns the area of the given Object (in object units). Returned is the area of the polygonal object, not the area of the primitive it might approximate.

This means that the area of a polygonal approximation of a sphere will be returned, not the exact area of the sphere.

ASIN

NumericType ASIN( NumericType Operand )

Returns the arc sine value (in radians) of the given Operand.

ATAN

NumericType ATAN( NumericType Operand )

Returns the arc tangent value (in radians) of the given Operand.

ATAN2

NumericType ATAN2( NumericType Operand1, NumericType Operand2 )

Returns the arc tangent value (in radians) of the given ratio: Operand1 / Operand2, over the whole circle.

COS

NumericType COS( NumericType Operand )

Returns the cosine value of the given Operand (in radians).

CLNTEXEC

NumericType CLNTEXEC( StringType ClientName )

Initiate communication channels to a client named ClientName. ClientName is executed by this function as a sub process and two communication channels are opened between the IRIT server and the new client, for read and write. See also CLNTREAD, CLNTWRITE, and CLNTCLOSE. if ClientName is an empty string, the user is provided with the new communication port to be used and the server blocks for the user to manualy executed the client after setting the proper IRIT_SERVER_HOST/PORT environment variables.

Example:

h1 = CLNTEXEC( "" ); h2 = CLNTEXEC( "nuldrvs -s-" );

executes two clients, one is named nuldrvs and the other one is prompted for by the user. As a result of the second invokation of CLNTEXEC, the user will be prompted with a message similar to,

Irit: Startup your program - I am waiting... setenv IRIT_SERVER_PORT 2182

and he/she will need to set the proper environment variable and execute their client manually.

CPOLY

NumericType CPOLY( PolygonType Object )

Returns the number of polygons in the given polygonal Object.

DSTPTLN

NumericType DSTPTLN( PointType Pt, PointType LineOrig, VectorType LineRay )

Returns the distance between a given point Pt and line LineOrig, LineRay. See also PTPTLN.

DSTPTPLN

NumericType DSTPTPLN( PointType Pt, PlaneType Plane )

Returns the distance between a given point Pt and plane Plane.

DSTLNLN

NumericType DSTLNLN( PointType Line1Orig, VectorType Line1Ray, PointType Line2Orig, VectorType Line2Ray )

Returns the distance between two lines defined by point LineiOrig and ray LineiRay. See also PTSLNLN.

EXP

NumericType EXP( NumericType Operand )

Returns the natural exponent value of the given Operand.

FLOOR

NumericType FLOOR( NumericType Operand )

Returns the largest integer not greater than Operand.

FMOD

NumericType FMOD( NumericType Operand, NumericType Mod )

Returns the floating point remainder of the division of Operand by Mod.

LN

NumericType LN( NumericType Operand )

Returns the natural logarithm value of the given Operand.

LOG

NumericType LOG( NumericType Operand )

Returns the base 10 logarithm value of the given Operand.

MESHSIZE

NumericType MESHSIZE( SurfaceType Srf, ConstantType Direction )

Returns the size of Srf's mesh in Direction, which is one of COL or ROW.

Example:

RSize = MESHSIZE( Sphere, ROW ); CSize = MESHSIZE( Sphere, COL );

POWER

NumericType POWER( NumericType Operand, NumericType Exp )

Returns Operand to the power of Exp.

RANDOM

NumericType RANDOM( NumericType Min, NumericType Max )

Returns a randomized value between Min and Max.

SIN

NumericType SIN( NumericType Operand )

Returns the sine value of the given Operand (in radians).

SIZEOF

NumericType SIZEOF( ListType List | PolygonType Poly | CurveType Crv | StringType Str )

Returns the length of a list if List, the number of polygons if Poly, the length of the control polygon if Crv, or the number of characters in string if Str. If, however, only one polygon is in Poly, it returns the number of vertices in that polygon.

Example:

len = SIZEOF( list( 1, 2, 3 ) ); numPolys = SIZEOF( axes ); numCtlpt = SIZEOF( circle( vector( 0, 0, 0 ), 1 ) );

will assign the value of 3 to the variable len, set numPolys to the number of polylines in the axes object, and set numCtlPt to 9, the number of control points in a circle.

SQRT

NumericType SQRT( NumericType Operand )

Returns the square root value of the given Operand.

TAN

NumericType TAN( NumericType Operand )

Returns the tangent value of the given Operand (in radians).

THISOBJ

NumericType THISOBJ( StringType Object )

Returns the object type of the given name of an Object. This can be one of the constants,

UNDEF_TYPE	VECTOR_TYPE	MATRIX_TYPE	SURFACE_TYPE
NUMERIC_TYPE	POINT_TYPE	POLY_TYPE	TRIMSRF_TYPE
STRING_TYPE	CTLPT_TYPE	CURVE_TYPE	TRIVAR_TYPE

This is also a way to ask if an object by a given name do exist (if the returned type is UNDEF_TYPE or not).

VOLUME

NumericType VOLUME( PolygonType Object )

Returns the volume of the given Object (in object units). It returns the volume of the polygonal object, not the volume of the object it might approximate.

This routine decomposes all non-convex polygons to convex ones as a side effect (see CONVEX).

GeometricType returning functions

ADAPISO

CurveType ADAPISO( SurfaceType Srf, NumericType Dir, NumericType Eps, NumericType FullIso, NumericType SinglePath )

Constructs a coverage to Srf using isocurve in the Dir direction, so that for any point p on surface Srf, there exists a point on one of the isocurves that is close to p within Eps. If FullIso, the extracted isocurves span the entire surface domain, otherwise they may span only a subset of the domain. If SinglePath, an approximation to a single path (Hamiltonian path) that visits all isocurves is constructed (not supported). See also COVERPT, COVERISO.

srf = sbezier( list( list( ctlpt( E3, -0.5, -1.0, 0.0 ), ctlpt( E3, 0.4, 0.0, 0.1 ), ctlpt( E3, -0.5, 1.0, 0.0 ) ), list( ctlpt( E3, 0.0, -0.7, 0.1 ), ctlpt( E3, 0.0, 0.0, 0.0 ), ctlpt( E3, 0.0, 0.7, -0.2 ) ), list( ctlpt( E3, 0.5, -1.0, 0.1 ), ctlpt( E3, -0.4, 0.0, 0.0 ), ctlpt( E3, 0.5, 1.0, -0.2 ) ) ) ); aiso = ADAPISO( srf, COL, 0.1, FALSE, FALSE );

Constructs an adaptive isocurve approximation with tolerance of 0.1 to surface srf in direction COL. Isocurves are allowed to span a subset of the surface domain. No single path is needed.

The SinglePath option is currently not supported.

ARC

CurveType ARC( VectorType StartPos, VectorType Center, VectorType EndPos )

Constructs an arc between the two end points StartPos and EndPos, centered at Center. Arc will always be less than 180 degrees, so the shortest circular path from StartPos to EndPos is selected. The case where StartPos, Center, and EndPos are collinear is illegal, since it attempts to define a 180 degrees arc. Arc is constructed as a single rational quadratic Bezier curve.

Example:

Arc1 = ARC( vector( 1.0, 0.0, 0.0 ), vector( 1.0, 1.0, 0.0 ), vector( 0.0, 1.0, 0.0 ) );

constructs a 90 degrees arc, tangent to both the X and Y axes at coordinate 1.

AOFFSET

CurveType AOFFSET( CurveType Crv, NumericType OffsetDistance, NumericType Epsilon, NumericType TrimLoops, NumericType BezInterp )

Computes an offset of OffsetDistance with globally bounded error (controlled by Epsilon). The smaller Epsilon is, the better the approximation to the offset. The bounded error is achieved by adaptive refinement of the Crv. If TrimLoops is TRUE or on, the regions of the object that self-intersect as a result of the offset operation are trimmed away. If BezInterp is TRUE, each curve's segment is interpolated instead of approximated.

Example:

OffCrv1 = AOFFSET( Crv, 0.5, 0.01, FALSE, FALSE ); OffCrv2 = AOFFSET( Crv, 0.5, 0.01, TRUE, FALSE );

computes an adaptive offset to Crv with OffsetDistance of 0.5 and Epsilon of 0.01 and trims the self-intersection loops in the second instrance. See also OFFSET, LOFFSET, and MOFFSET.

BBOX

ListType BBOX( GeometricTreeType Geom )

Given a (tree of ) geometry, Geom computes its bounding box and return it as a list of six numbers: XMin/Max, YMin/Max, ZMin/Max, in this order.

Example:

B1 = BBOX( axes );

BOOLONE

SurfaceType BOOLONE( CurveType Crv )

Given a closed curve, the curve is subdivided into four segments equally spaced in the parametric space that are fed into BOOLSUM. Useful if a surface should "fill" the area enclosed by a closed curve.

Example:

Srf = BOOLONE( circle( vector( 0.0, 0.0, 0.0 ), 1.0 ) );

Creates a disk surface containing the area enclosed by the unit circle.

BOOLSUM

SurfaceType BOOLSUM( CurveType Crv1, CurveType Crv2, CurveType Crv3, CurveType Crv4 )

Construct a surface using the provided four curves as its four boundary curves. Curves do not have to have the same order or type, and will be promoted to their least common denominator. The end points of the four curves should match as follows:

Crv1 start point,	to Crv3 start point.
Crv1 end point,	to Crv4 start point.
Crv2 start point,	to Crv3 end point.
Crv2 end point,	to Crv4 end point.

where Crv1 and Crv2 are the two boundaries in one parametric direction, and Crv3 and Crv4 are the two boundaries in the other parametric direction.

Example:

Cbzr1 = cbezier( list( ctlpt( E3, 0.1, 0.1, 0.1 ), ctlpt( E3, 0.0, 0.5, 1.0 ), ctlpt( E3, 0.4, 1.0, 0.4 ) ) ); Cbzr2 = cbezier( list( ctlpt( E3, 1.0, 0.2, 0.2 ), ctlpt( E3, 1.0, 0.5, -1.0 ), ctlpt( E3, 1.0, 1.0, 0.3 ) ) ); Cbsp3 = cbspline( 4, list( ctlpt( E3, 0.1, 0.1, 0.1 ), ctlpt( E3, 0.25, 0.0, -1.0 ), ctlpt( E3, 0.5, 0.0, 2.0 ), ctlpt( E3, 0.75, 0.0, -1.0 ), ctlpt( E3, 1.0, 0.2, 0.2 ) ), list( KV_OPEN ) ); Cbsp4 = cbspline( 4, list( ctlpt( E3, 0.4, 1.0, 0.4 ), ctlpt( E3, 0.25, 1.0, 1.0 ), ctlpt( E3, 0.5, 1.0, -2.0 ), ctlpt( E3, 0.75, 1.0, 1.0 ), ctlpt( E3, 1.0, 1.0, 0.3 ) ), list( KV_OPEN ) ); Srf = BOOLSUM( Cbzr1, Cbzr2, Cbsp3, Cbsp4 );

BOX

PolygonType BOX( VectorType Point, NumericType Dx, NumericType Dy, NumericType Dz )

Creates a BOX polygonal object, whose boundary is coplanar with the XY, XZ, and YZ planes. The BOX is defined by Point as base position, and Dx, Dy, Dz as BOX dimensions. Negative dimensions are allowed.

Example:

B = BOX( vector( 0, 0, 0 ), 1, 1, 1);

creates a unit cube from 0 to 1 in all axes.

BZR2BSP

CurveType BZR2BSP( CurveType Crv ) or SurfaceType BZR2BSP( SurfaceType Srf )

Creates a Bspline curve or a Bspline surface from the given Bezier curve or Bezier surface. The Bspline curve or surface is assigned open end knot vector(s) with no interior knots, in the parametric domain of zero to one.

Example:

BspSrf = BZR2BSP( BzrSrf );

BSP2BZR

CurveType | ListType BSP2BZR( CurveType Crv ) or SurfaceType | ListType BSP2BZR( SurfaceType Srf )

Creates Bezier curve(s) or surface(s) from a given Bspline curve or a Bspline surface. The Bspline input is subdivided at all internal knots to create Bezier curves or surfaces. Therefore, if the input Bspline does have internal knots, a list of Bezier curves or surfaces is returned. Otherwise, a single Bezier curve or surface is returned.

Example:

BzrCirc = BSP2BZR( circle( vector( 0.0, 0.0, 0.0 ), 1.0 ) );

would subdivide the unit circle into four 90 degrees Bezier arcs returned in a list.

CBEZIER

CurveType CBEZIER( ListType CtlPtList )

Creates a Bezier curve out of the provided control point list. CtlPtList is a list of control points, all of which must be of type (E1-E5, P1-P5), or regular PointType defining the curve's control polygon. Curve's point type will be of a space which is the union of the spaces of all points.

Example:

s45 = sin(pi / 4); Arc90 = CBEZIER( list( ctlpt( P2, 1.0, 0.0, 1.0 ), ctlpt( P2, s45, s45, s45 ), ctlpt( P1, 1.0, 1.0 ) ) );

constructs an arc of 90 degrees as a rational quadratic Bezier curve.

CBISECTOR

{ CurveType | SurfaceType } CBISECTOR( CurveType Crv, NumericType UseNrmlTan, NumericType Tolerance, NumericType NumerImprove, NumericType SameNormal ) or { CurveType | SurfaceType } CBISECTOR( ListType TwoCrvs, NumericType UseNrmlTan, NumericType Tolerance, NumericType NumerImprove, NumericType SameNormal ) or SurfaceType CBISECTOR( ListType CrvPt, NumericType UseNrmlTan, NumericType Tolerance, NumericType NumerImprove, NumericType SameNormal )

Computes the self bisector curve(s) for Crv or the bisector(s) of TwoCrvs or the bisector of a curve and a point, CrvPt. If UseNrmlTan is equal to zero or one, the bisector curves are computed using angular equality of tangents or normals of the given curve or curves, respectively. Tolerance controls the accuracy of the computation, with 10 as a good starting value. If Tolerance is negative, NumerImprove can be either TRUE or FALSE, allowing or disabling a final numerical imrpovement stage. SameNormal can also assume a TRUE or FALSE value, selecting only opposite facing normals, if TRUE.

Nevertheless, if UseNrmlTan equals -1, the bisector surface of the given two curves TwoCrvs or a curve and a point CrvPt is computed analytically. For two 3-space curves this is indeed the bisector surface while for two 2-space curves, the zero set of the returned surface provides the bisector curve. If UseNrmlTan equals -1, then Tolerance, NumerImprove, and SameNormal are all ignored.

If UseNrmlTan equals -2 and the two curves are E2, a different bisector surface is computed whose zero set provides the bisector curve. If UseNrmlTan equals -2, then Tolerance, NumerImprove, and SameNormal are all ignored.

If UseNrmlTan equals -3 and the two curves are 3-space curves, an alpha bisector surface is computed, whose ratio is set via Tolerance relative to NumerImprove, such that the sum of Tolerance + NumerImprove is not zero. If UseNrmlTan equals -3, then SameNormal is ignored. See also SBISECTOR.

Example:

c1 = cbezier( list( ctlpt( E2, -0.5, -0.2 ), ctlpt( E2, 0.0, -0.2 ), ctlpt( E2, 0.6, 0.6 ) ) ); c2 = cbezier( list( ctlpt( E2, 0.3, -0.7 ), ctlpt( E2, -0.2, -0.7 ), ctlpt( E2, 0.7, 0.6 ) ) ); BisectCrvs = CBISECTOR( list( c1, c2 ), 0, 12, true, false ); All = list( c1, c2, BisectCrvs ); interact( list( All, view_mat2d ) ); c1 = creparam( pcircle( vector( 0.0, 0.0, 0.0 ), 1 ), 0, 1 ); c2 = cbezier( list( ctlpt( E3, -1.0, 0.0, 1.0 ), ctlpt( E3, 1.0, 0.0, -1.0 ) ) ); BisectSrf = CBISECTOR( list( c1, c2 ), -1, -1, true, false ); interact( list( c1, c2, BisectSrf ) );

computes two bisectors, one for to planar curves as a set of bisector curves, and the other as a bisector surfce of a Z parallel line and a circle in the XY plane.

CBSPLINE

CurveType CBSPLINE( NumericType Order, ListType CtlPtList, ListType KnotVector )

Creates a Bspline curve out of the provided control point list, the knot vector, and the specified order. CtlPtList is a list of control points, all of which must be of type (E1-E5, P1-P5), or regular PointType defining the curve's control polygon. Curve's point type will be of a space which is the union of the spaces of all points. The length of the KnotVector must be equal to the number of control points in CtlPtList plus the Order. If, however, the length of the knot vector is equal to #CtlPtList + Order + Order - 1 the curve is assumed periodic. The knot vector list may be specified as either list( KV_OPEN ) or list( KV_FLOAT ) or list( KV_PERIODIC ) in which a uniform open, uniform floating or uniform periodic knot vector with the appropriate length is automatically constructed.

Example:

s45 = sin(pi / 4); HalfCirc = CBSPLINE( 3, list( ctlpt( P3, 1.0, 1.0, 0.0, 0.0 ), ctlpt( P3, s45, s45, s45, 0.0 ), ctlpt( P3, 1.0, 0.0, 1.0, 0.0 ), ctlpt( P3, s45, -s45, s45, 0.0 ), ctlpt( P3, 1.0, -1.0, 0.0, 0.0 ) ), list( 0, 0, 0, 1, 1, 2, 2, 2 ) );

constructs an arc of 180 degrees in the XZ plane as a rational quadratic Bspline curve.

Example:

c = CBSPLINE( 4, list( ctlpt( E2, 0.5, 0.5 ), ctlpt( E2, -0.5, 0.5 ), ctlpt( E2, -0.5, -0.5 ), ctlpt( E2, 0.5, -0.5 ) ), list( KV_PERIODIC ) ); color( c, red ); viewobj( c ); c1 = cregion( c, 3, 4 ); color( c1, green ); c2 = cregion( c, 4, 5 ); color( c2, yellow ); c3 = cregion( c, 5, 6 ); color( c3, cyan ); c4 = cregion( c, 6, 7 ); color( c3, magenta ); viewobj( list( c1, c2, c3, c4 ) );

creates a periodic curve and extracts its four polynomial domains as four open end Bspline curves.

CCINTER

ListType CCINTER( CurveType Crv1, CurveType Crv2, NumericType Epsilon, NumericType SelfInter ) or SurfaceType CCINTER( CurveType Crv1, CurveType Crv2, NumericType Epsilon, NumericType SelfInter )

Computes the intersection point(s) of Crv1 and Crv2 in the XY plane. Since this computation involves numeric operations, Epsilon controls the accuracy of the parametric values of the result. It returns a list of PointTypes, each containing the parameter of Crv1 in the X coordinate, and the parameter of Crv2 in the Y coordinate. If, however, Epsilon is negative, a scalar field surface representing the square of the distance function is returned instead. If SelfInter is TRUE, Crv1 and Crv2 can be the same curve, and self-intersection points are searched instead.

Example:

crv1 = cbspline( 3, list( ctlpt( E2, 0, 0 ), ctlpt( E2, 0, 0.5 ), ctlpt( E2, 0.5, 0.7 ), ctlpt( E2, 1, 1 ) ), list( KV_OPEN ) ); crv2 = cbspline( 3, list( ctlpt( E2, 1, 0 ), ctlpt( E2, 0.7, 0.25 ), ctlpt( E2, 0.3, 0.5 ), ctlpt( E2, 0, 1 ) ), list( KV_OPEN ) ); inter_pts = CCINTER( crv1, crv2, 0.0001, FALSE );

Computes the parameter values of the intersection point of crv1 and crv2 to a tolerance of 0.0001.

CCRVTR

NumericType CCRVTR( CurveType Crv, NumericType Epsilon ) or CurveType CCRVTR( CurveType Crv, NumericType Epsilon )

Computes the extreme curvature points on Crv in the XY plane. This set includes not only points of maximum (convexity) and mimumum (concavity) curvature, but also points of zero curvature such as inflection points. Since this operation is partially numeric, Epsilon is used to set the needed accuracy. It returns the parameter value(s) of the location(s) with extreme curvature along the Crv. If, however, Epsilon is negative, the curvature scalar field curve is returned as a two dimensional rational vector field curve, for which the first dimension is equal to the parameter, and the second is the curvature value at that parameter.

   This function computes the curvature scalar field for planar curves as,

          x' y'' - x'' y'

   k(t) = ----------------

               2     2  3/2

           ( x'  + y'  )

 and computes kN for three dimensional curves as the following vector field,

                                  C' x C''     C'    (C' x C'') x C'

   k(t) N(t) = K(t) B(t) x T(t) = -------- x ----- = ---------------

                                        3    | C' |            4

                                   | C'|                 | C' |

The extremum values are extracted from the computed curvature field. This curvature field is a high order curve, even if the input geometry is of low order. This is especially true for rational curves, for which the quotient rule for differentiation is used and almost doubles the degree in every differentiation.

See also CZEROS, CEXTREMES, and SCRVTR.

Example:

crv = cbezier( list( ctlpt( E2, -1.0, 0.5 ), ctlpt( E2, -0.5, -2.0 ), ctlpt( E2, 0.0, 1.0 ), ctlpt( E2, 1.0, 0.0 ) ) ) * rotz( 30 ); crvtr = CCRVTR( crv, 0.001 ); pt_crvtr = nil(); pt = nil(); for ( i = 1, 1, sizeof( crvtr ), ( pt = ceval( crv, nth( crvtr, i ) ) ): snoc( pt, pt_crvtr ) ); interact( list( crv, pt_crvtr ) );

finds the extreme curvature points in Crv and displays them all with the curve.

CDERIVE

CurveType CDERIVE( CurveType Curve )

Returns a vector field curve representing the differentiated curve, also known as the Hodograph curve.

Example:

Circ = circle( vector( 0.0, 0.0, 0.0 ), 1.0 ); Hodograph = CDERIVE( Circ );

CDIVIDE

ListType CDIVIDE( CurveType Curve, NumericType Param )

Subdivides a curve into two sub-curves at the specified parameter value. Curve can be either a Bspline curve in which Param must be within the Curve's parametric domain, or a Bezier curve in which Param can be arbitrary, extrapolating if not in the range of zero to one.

Returns a list of the two sub-curves. The individual curves may be extracted from the list using the NTH command.

Example:

CrvLst = CDIVIDE( Crv, 1.3 ); Crv1 = nth( CrvLst, 1 ); Crv2 = nth( CrvLst, 2 );

subdivides the curve Crv at the parameter value of 0.5.

CEDITPT

CurveType CEDITPT( CurveType Curve, CtlPtType CtlPt, NumericType Index )

Provides a simple mechanism to manually modify a single control point number Index (base count is 0) in Curve, by substituting CtlPt instead. CtlPt must have the same point type as the control points of the Curve. Original curve Curve is not modified.

Example:

CPt = ctlpt( E3, 1, 2, 3 ); NewCrv = CEDITPT( Curve, CPt, 1 );

constructs a NewCrv with the second control point of Curve being CPt.

CENVOFF

SurfaceType CENVOFF( CurveType Curve, NumericType Height, NumericType Tolerance ) or ListType CENVOFF( CurveType Curve, NumericType Height, NumericType Tolerance )

Computes the offset envelope of a given planar curve Curve. The offset envelope is the envelope of cones with apex on point on Curve in the Z direction. Height is the height of the cone which also equals the offset distance or the width of the cones. Tolerance controls the accuracy of the offset approximation.

If Curve is closed, two surfaces are created in the offset envelope, one for the inside and one for the outside. If Curve is open, a single envelope offset surface is computed wrapping around both sides.

Example:

c1 = cbezier( list( ctlpt( E2, -0.8, 0.0 ), ctlpt( E2, -0.2, 1.0 ), ctlpt( E2, 0.2, 0.0 ), ctlpt( E2, 0.8, 0.6 ) ) ); s1 = CENVOFF( c1, 0.5, 0.01 );

Computes an envelope offset surface for a cubic Bezier curve c1 of Height of 0.5 and Tolerance of 0.01.

CEVAL

CtlPtType CEVAL( CurveType Curve, NumericType Param )

Evaluates the provided Curve at the given Param value. Param should be in the curve's parametric domain if Curve is a Bspline curve, or between zero and one if Curve is a Bezier curve. The returned control point has the same point type as the control points of the Curve.

Example:

CPt = CEVAL( Crv, 0.25 );

evaluates Crv at the parameter value of 0.25.

CEXTREMES

ListType CEXTREMES( CurveType Crv, NumericType Epsilon, NumericType Axis )

Computes the extreme set of the given Crv in the given axis (1 for X, 2 for Y, 3 for Z). Since this computation is numeric, an Epsilon is also required to specify the desired tolerance. It returns a list of all the parameter values (NumericType) in which the curve takes an extreme value.

Example:

extremes = CEXTREMES( Crv, 0.0001, 1 );

Computes the extreme set of curve crv, in the X axis, with error tolerance of 0.0001. See also CZERO.

CINFLECT

ListType CINFLECT( CurveType Crv, NumericType Epsilon ) or CurveType CINFLECT( CurveType Crv, NumericType Epsilon )

Computes the inflection points of Crv in the XY plane. Since this computation is numeric, an Epsilon is also required to specify the desired tolerance. It returns a list of all the parameter values (NumericType) in which the curve has an inflection point. If, however, Epsilon is negative, a scalar field curve representing the sign of the curvature of the curve is returned instead.

 The sign of curvature scalar field is equal to

       s(t) = x' y'' - x'' y'

Example:

inflect = CINFLECT( crv, 0.001 ); pt_inflect = nil(); pt = nil(); for ( i = 1, 1, sizeof( inflect ), pt = ceval( crv, nth( inflect, i ) ): snoc( pt, pt_inflect ) ); interact( list( axes, crv, pt_inflect ), 0);

Computes the set of inflection points of curve crv with error tolerance of 0.001. This set is then scanned in a loop and evaluated to the curve's locations which are then displayed with the crv. See also CZEROS, CEXTREMES, and CCRVTR.

CINTERP

CurveType CINTERP( ListType PtList, NumericType Order, NumericType Size, ConstantType Param, NumericType Periodic)

Computes a Bspline polynomial curve that interpolates or approximates the list of points in PtList. The Bspline curve will have order Order and Size control points, and will be periodic if periodic is none zero. The knots will be spaced according to Param which can be one of PARAM_UNIFORM, PARAM_CHORD or PARAM_CENTRIP. The former prescribes a uniform knot sequence and the latters specify knot spacing according to the chord length and a square root of the chord length. A periodic curve will be coerced to have PARAM_UNIFORM knot sequence. Use of Periodic end conditions can create cases with degenerated linear systems (determinant equal zero). Increase or decrease of the Order of the Bspline by one will resolve the problem. All points in PtList must be of type (E1-E5, P1-P5) control point, or regular PointType. If Size is equal to the number of points in PtList the resulting curve will interpolate the data set. Otherwise, if Size is less than the number of points in PtList the point data set will be least square approximated. In no time can Size be lower than Order. Size of zero forces interpolation by setting Size to the data set size. All interior knots will be distinct preserving maximal continuity. The resulting Bspline curve will have open end conditions.

Example:

pl = nil(); for ( x = 0, 1, 100, snoc(point(cos(x / 5), sin(x / 5), x / 50 - 1), pl) ); c = CINTERP( pl, 3, 21, PARAM_UNIFORM );

Samples a helical curve at 100 points and least square fit a quadratic Bspline curve with 21 point to the data set. The curve will have a uniform knot spacing.

CIRCLE

CurveType CIRCLE( VectorType Center, NumericType Radius )

Constructs a circle at the specified Center with the specified Radius. The returned circle is a Bspline curve of four piecewise Bezier 90 degree arcs. The construced circle is always parallel to the XY plane. Use the linear transformation routines to place the circle in the appropriate orientation and location.

CIRCPOLY

PolygonType CIRCPOLY( VectorType Normal, VectorType Trans, NumericType Radius )

Defines a circular polygon in a plane perpendicular to Normal that contains the Trans point. Constructed polygon is centered at Trans. RESOLUTION vertices will be defined with Radius from distance from Trans.

Alternative ways to construct a polygon are manual construction of the vertices using POLY, or the construction of a flat ruled surface using RULEDSRF.

CLNTREAD

AnyType CLNTREAD( NumericType Handler, NumericType Block )

Reads one object from a communication channel of a client. Handler contains the index of the communication channel opened via CLNTEXEC. If no data is available in the communication channel, this function will block for at most Block millisecond until data is found or timeout occurs. In the latter, a single StringType object is returned with the content of "no data (timeout)". See also CLNTWRITE, CLNTCLOSE, and CLNTEXEC.

Example:

h2 = clntexec( "xmtdrvs -s-" ); . . Model = CLNTREAD( h2 ); . . clntclose( h2,TRUE );

reads one object from client through communication channel h2 and save it in variable Model.

CMESH

CurveType CMESH( SurfaceType Srf, ConstantType Direction, NumericType Index )

Returns a single ROW or COLumn as specified by the Direction and Index (base count is 0) of the control mesh of surface Srf.

The returned curve will have the same knot vector as Srf in the appropriate direction. See also CSURFACE.

This curve is not necessarily in the surface Srf.

Example:

Crv = CMESH( Srf, COL, 0 );

extracts the first column of surface Srf as a curve. See also CSURFACE.

CMORPH

CurveType CMORPH( CurveType Crv1, CurveType Crv2, NumericType Method, NumericType Blend ) or ListType CMORPH( CurveType Crv1, CurveType Crv2, NumericType Method, NumericType Blend )

Creates a new curve which is a metamorph of the two given curves. The two given curves must be compatible (see FFCOMPAT) before this blend is invoked. Very useful if a sequence that "morphs" one curve to another is to be created. Several methods of metamorphosis are supported according to the value of Method,

0	Simple convex blend.
1	Corner/Edge cutting scheme, scaled to same curve length.
2	Corner/Edge cutting scheme, scaled to same bounding box.
3	Same as 1 but with filtering out of tangencies.
4	Same as 2 but with filtering out of tangencies.
5	Multiresolution decompsition based metamorphosis. See CMULTRES.

In Method 1, Blend is a number between zero (Crv1) and one (Crv2) defining the similarity to Crv1 and Crv2, respectively. A single curve is returned.

In Methods 2 to 5, Blend is a step size for the metamorphosis operation and a whole list describing the entire metamorphosis operation is returned.

Examples:

for ( i = 0, 1, 300, c = CMORPH( crv1a, crv1b, 0, i / 300.0 ): color( c, yellow ): viewobj( c ) ); crvs = CMORPH( crv1a, crv1b, 2, 0.003 ); snoc( crv1b, crvs ); for ( i = 1, 1, sizeof( crvs ), c = nth( crvs, i ): color( c, yellow ): viewobj( c ) ); Turtle2 = ffmatch( Wolf, Turtle, 20, 100, 2, false, 2 ); ffcompat( Wolf, Turtle2 ); for ( i = 0, 1, 25, c = CMORPH( Wolf, Turtle2, 0, i / 25 ): color( c, yellow ): viewobj( c ) );

creates three metamorphosis animation sequences, one that is based on a convex blend, one that is based on corner/edge cutting scheme. See alost SMORPH, and a third that employs FFMATCH.

CMULTIRES

ListType CMULTIRES( CurveType Crv, NumericType Discont )

Computes a multiresolution decomposition of curve Crv using least squares approximation. The resulting list of curves describes an hierarchy of curves in linear subspaces of the space Crv was in that can be sum algebraically to form Crv. Each of the curves in the hierarchy is a least squares approximation of Crv in the subspace it is defined in. Discont is a boolean flat that controls the way tangent discontinuities are computed throughout the decomposition.

Example:

MRCrv = CMULTIRES( Animal, false ); sum = nth( MRCrv, 1 ); MRCrvs = list( sum * tx( 3.0 ) ); for ( ( i = 2 ), 1, sizeof( MRCrv ), sum = symbsum( sum, nth( MRCrv, i ) ): snoc( sum * tx( ( 3 - i ) * 1.5 ), MRCrvs ) ); All = MRCrvs * sc ( 0.25 ); view( All, on );

Computes a multiresolution decomposition to curve CrossSec as MRCrv and display all the decomposition levels by summing them all up. The use of none as on object name allows one to display an object in the display device without replacing the previous object in the display device, carrying the same name.

creates two metamorphosis animation sequences, one that is based on a convex blend and one that is based on corner/edge cutting scheme.

CNORMAL

VectorType CNORMAL( CurveType Crv, NumericType TParam )

Computes the normal vector to curve Crv at the parameter values TParam. The returned vector has a unit length.

Example:

Normal = CNORMAL( Crv, 0.5 );

computes the normal to Crv at the parameter value 0.5. See also CNRMLCRV.

CNRMLCRV

CurveType CNRMLCRV( CurveType Crv )

Symbolically computes a vector field curve representing the non-normalized normals of the given curve. That is a normal vector filed, evaluated at t, provides a vector in the direction of the normal of the original curve at t. The normal curve computed is in fact equal to kN where k is the curvature of Crv and N is its normal.

Example:

NrmlCrv = CNRMLCRV( Crv );

CNVXHULL

PolygonType CNVXHULL( PolygonType Poly | PolylineType Poly ); or CurveType CNVXHULL( CurveType Crv );

Computes the convex hull of the given set of Poly or Curve. For curves, the result might be partial if the curve is not closed or periodic. See also CRV2TANS and CRVPTTAN.

Example:

Pts1 = nil(); Pts2 = nil(); for ( i = 0, 1, 7, R = 0.2 + fmod( i, 2 ) / 2: Pt = ctlpt( E2, R * cos( i * 2 * pi / 8 ), R * sin( i * 2 * pi / 8 ) ): snoc( Pt, Pts1 ): snoc( coerce( Pt, point_type ), Pts2 ) ); Crv = coerce( cbspline( 4, Pts1, list( KV_PERIODIC ) ), KV_OPEN ); CHPts = CNVXHULL( poly( Pts2, 0 ), 0 ); CHCrv = CNVXHULL( Crv, 0 );

computes the convex hull of a given control polygon and freeform curve.

COERCE

AnyType COERCE( AnyType Object, ConstantType NewType )

Provides a coercion mechanism between different objects or object types. PointType, VectorType, PlaneType, CtlPtType can be all coerced to each other by using the NewType of POINT_TYPE, VECTOR_TYPE, PLANE_TYPE, or one of E1-E5, P1-P5 (CtlPtType). Similarly, CurveType, SurfaceType, TrimSrfType, TriSrfType, and TrivarType can all be coerced to hold different CtlPtType of control points, or even different open end conditions from KV_PERIODIC to KV_FLOAT to KV_OPEN. If a scalar (E1 or P1) curve is coerced to E2 or P2 curve or a scalar (E1 or P1) surface is coerced to E3 or P3 surface, the Y (YZ) coordinate(s) is (are) updated to hold the parametric domain of the curve (surface). That is X = U (Y = V).

Example:

CrvE2 = COERCE( Crv, E2 );

coerce Crv to a new curve that will have an E2 CtlPtType control points. Coerction of a projective curve (P1-P5) to a Euclidean curve (E1-E5) does not preseve the shape of the curve.

COMPOSE

CurveType COMPOSE( CurveType Crv1, CurveType Crv2 ) or CurveType COMPOSE( SurfaceType Srf, CurveType Crv )

Symbolically compute the composition curve Crv1(Crv2(t)) or Srf(Crv(t)). In Crv1(Crv2(t), Crv1 can be any curve while Crv2 must be a one-dimensional curve that is either E1 or P1. In Srf(Crv(t)), Srf can be any surface, while Crv must be a two-dimensional curve, that is either E2 or P2. Both Crv2 in the curve's composition, and Crv is the surface's composition must be contained in the curve or surface parametric domain.

Example:

srf = sbezier( list( list( ctlpt( E3, 0.0, 0.0, 0.0 ), ctlpt( E3, 0.0, 0.5, 1.0 ), ctlpt( E3, 0.0, 1.0, 0.0 ) ), list( ctlpt( E3, 0.5, 0.0, 1.0 ), ctlpt( E3, 0.5, 0.5, 0.0 ), ctlpt( E3, 0.5, 1.0, 1.0 ) ), list( ctlpt( E3, 1.0, 0.0, 1.0 ), ctlpt( E3, 1.0, 0.5, 0.0 ), ctlpt( E3, 1.0, 1.0, 1.0 ) ) ) ); crv = coerce( circle( vector( 0.0, 0.0, 1.0 ), 0.4 ), p2 ) * trans( vector( 0.5, 0.5, 0.0 ) ); comp_crv = COMPOSE( srf, crv );

composes a circle Crv to be on the surface Srf.

CON2

PolygonType CON2( VectorType Center, VectorType Direction, NumericType Radius1, NumericType Radius2 )

Creates a truncated CONE geometric object, defined by Center as the center of the main base of the CONE, Direction as both the CONE's axis and the length of CONE, and the two radii Radius1/2 of the two bases of the CONE.

Unlike the regular cone (CONE) constructor which has inherited discontinuities in its generated normals at the apex, CON2 can be used to form a (truncated) cone with continuous normals. See RESOLUTION for the accuracy of the CON2 approximation as a polygonal model. See also CONE. See IRITSTATE's "PrimRatSrfs" and "PrimRatSrfs" state variables.

Example:

Cone2 = CON2( vector( 0, 0, -1 ), vector( 0, 0, 4 ), 2, 1 );

constructs a truncated cone with bases parallel to the XY plane at Z = -1 and Z = 3, and with radii of 2 and 1 respectively.

CONE

PolygonType CONE( VectorType Center, VectorType Direction, NumericType Radius )

Creates a CONE geometric object, defined by Center as the center of the base of the CONE, Direction as the CONE's axis and height, and Radius as the radius of the base of the CONE. See RESOLUTION for accuracy of the CONE approximation as a polygonal model.

Example:

Cone1 = CONE( vector( 0, 0, 0 ), vector( 1, 1, 1 ), 1 );

constructs a cone based in an XY parallel plane, centered at the origin with radius 1 and with tilted apex at ( 1, 1, 1 ).

See IRITSTATE's "PrimRatSrfs" and "PrimRatSrfs" state variables. See also CON2.

CONTOUR

PolygonType CONTOUR( SurfaceType ContouredSrf, PlaneType ContourPlane ) or PolygonType CONTOUR( SurfaceType ContouredSrf, PlaneType ContourPlane, SurfaceType MappedSrf )

Contours surface ContouredSrf by intersecting plane ContourPlane with a polygonal approximating of ContouredSrf with resolution set via variable RESOLUTION (10 is a good starting resolutionvalue, 20 is extreme). If ContouredSrf is a scalar field surface of type E1 or P1 and MappedSrf is provided, ContouredSrf is contoured above its parametric domain (U is X, V is Y) and the resulting parametric curve is composed with MappedSrf to yield the returned value.

Example:

resolution = 20; nglass = snrmlsrf( glass ) * vector( 1, 1, 1 ); sils = contour( nglass, plane( 1, 0, 0, 0 ), glass ); color( sils, cyan ); attrib( sils, "dwidth", 4 ); view( list( axes, glass, sils ), on );

Computes the normal field of the surface glass, project it onto the viewing direction of (1, 1, 1) and contour the resulting scalar field with the plane X = 0, to extract the silhouette curves from viewing direction (1, 1, 1).

CONVEX

PolygonType CONVEX( PolygonType Object ) or ListType CONVEX( ListType Object )

Converts non-convex polygons in Object, into convex ones. New vertices are introduced into the polygonal data during this process. The Boolean operations require the input to have convex polygons only (although it may return non convex polygons...) and it automatically converts non-convex input polygons to convex ones, using this same routine.

However, some external tools (like irit2ray and poly3d-h) require convex polygons. This function must be used on the objects to guarantee that only convex polygons are saved into data files for these external tools.

Example:

CnvxObj = CONVEX( Obj ); save( "data", CnvxObj );

converts non-convex polygons into convex ones, so that the data file can be used by external tools requiring convex polygons.

COORD

AnyType COORD( AnyType Object, NumericType Index )

Extracts an element from a given Object, at index Index. From a PointType, VectorType, PlaneType, CtlPtType and MatrixType, a NumericType is returned with Index 0 for the X axis, 1 for the Y axis etc. Index 0 denotes the weight of CtlPtType. For a PolygonType that contains more than one polygon, the Indexth polygon is returned. For a PolygonType that contains a single Polygon, the Indexth vertex is returned. For a CurveType or a SurfaceType, the Indexth CtlPtType is returned. For a ListType, COORD behaves like NTH and returns the Indexth object in the list. For a StringType, the Indexth character is returned as its ASCII numeric code.

Example:

a = vector( 1, 2, 3 ); vector( COORD( a, 0 ), COORD( a, 1 ), COORD( a, 2 ) ); a = ctlpt( P2, 6, 7, 8, 9 ); ctlpt( P3, coord( a, 0 ), coord( a, 1 ), coord( a, 2 ), coord( a, 3 ) ); a = plane( 10, 11, 12, 13 ); plane( COORD( a, 0 ), COORD( a, 1 ), COORD( a, 2 ), COORD( a, 3 ) );

constructs a vector/ctlpt/plane and reconstructs it by extracting the constructed scalar components of the objects using COORD.

See also COERCE.

COVERISO

CurveType COVERISO( TrivarType TV, NumericType NewOfStrokes, NumericType StrokeType, PointType MinMaxPwrLen, NumericType StepSize, NumericType IsoVal )

Computes a coverage for an iso surface of a trivariate function TV, using curves. NewOfStrokes strokes are distributed on the iso surface with length that is set via MinMaxPwrLen. MinMaxPwrLen is a triplet of the form (Min, Max, Power) that determines the length of the strokes as,

        Avg = (Max + Min) / 2,             Dev = (Max - Min) / 2

        Length = Avg + Dev * Random(0, 1) ^ Pwr.

StepSize controls the steps size of the piecewise linear approximation formed and should be typically smaller than Min. StrokeType can be one of,

1	Draw strokes along minimal principal curvature.
2	Draw strokes along maximal principal curvature.
3	Draw strokes along both principal curvatures.
4	Draw strokes along constant X planes.
5	Draw strokes along constant Y planes.
6	Draw strokes along constant Z planes.

Finally, IsoVal controls the constant value of the iso surface level. See also ADAPISO, COVERPT, MRCHCUBE, TVLOAD.

Example:

IsoVal = 0.12; Cover = CoverIso( ThreeCyls, 500, 1, vector( 3, 10, 1.0 ), 0.2, IsoVal );

draws 500 strokes on the iso surface of trivariate ThreeCyls at iso value IsoVal and step size of 0.2. Strokes are drawn in length of 3 to 10 along line of curvatures of minimal curvature.

COVERPT

PolygonType COVERPT( PolygonType Model, NumericType NumOfPts, VectorType ViewDir )

Computes a uniform point distribution on the given polygonal Model. Approximately NumOfPts points are uniformly distributed on the Model's surface, provided ViewDir is the zero vector. If ViewDir is a non zero vector, the distribution is made to be uniform from this given viewing direction. In all cases, NumOfPts is an upper bound the the real number of distributed points, which will be in the same order.

See also ADAPISO, COVERISO, FFPTDIST.

Example:

Pts1 = CoverPt( solid1, 1000, vector( 0, 0, 0 ) ); Pts2 = CoverPt( solid1, 1000, vector( 0, 0, 1 ) );

Computes two uniform distributions of 1000 points on Solid1. Pts1 is uniform in three space, while Pts2 which is viewed uniform from the Z direction.

CRAISE

CurveType CRAISE( CurveType Curve, NumericType NewOrder )

Raise Curve to the NewOrder Order specified.

Example:

Crv = cbezier( list( ctlpt( E2, -0.7, 0.3 ), ctlpt( E2, 0.0, 1.0 ), ctlpt( E2, 0.7, 0.0 ) ) ); Crv2 = CRAISE( Crv, 5 );

raises the 90 degrees corner Bezier curve Crv to be a quartic.

CREFINE

CurveType CREFINE( CurveType Curve, NumericType Replace, ListType KnotList )

Provides the ability to Replace a knot vector of Curve, or refine it. KnotList is a list of knots to refine Curve at. All knots should be contained in the parametric domain of the Curve. If the knot vector is replaced, the length of KnotList should be identical to the length of the original knot vector of the Curve. If Curve is a Bezier curve, it is automatically promoted to be a Bspline curve.

Example:

Crv2 = CREFINE( Crv, FALSE, list( 0.25, 0.5, 0.75 ) );

refines Crv and adds three new knots at 0.25, 0.5, and 0.75.

CREGION

CurveType CREGION( CurveType Curve, NumericType MinParam, NumericType MaxParam )

Extracts a region from Curve between MinParam and MaxParam. Both MinParam and MaxParam should be contained in the parametric domain of the Curve.

Example:

SubCrv = CREGION( Crv, 0.3, 0.6 );

extracts the region from Crv from the parameter value 0.3 to the parameter value 0.6.

CREPARAM

CurveType CREPARAM( CurveType Curve, NumericType MinParam, NumericType MaxParam )

Reparametrize Curve over a new domain from MinParam to MaxParam. This operation does not affect the geometry of the curve and only affine transforms its knot vector. A Bezier curve will automatically be promoted into a Bspline curve by this function.

Example:

arc1 = arc( vector( 0.0, 0.0, 0.0 ), vector( 0.5, 2.0, 0.0 ), vector( 1.0, 0.0, 0.0 ) ); crv1 = arc( vector( 1.0, 0.0, 0.75 ), vector( 0.75, 0.0, 0.7 ), vector( 0.5, 0.0, 0.85 ) ) + arc( vector( 0.5, 0.0, 0.75 ), vector( 0.75, 0.0, 0.8 ), vector( 1.0, 0.0, 0.65 ) ); arc1 = CREPARAM( arc1, 0, 10 ); crv1 = CREPARAM( crv1, 0, 10 );

Sets the domain of the given two curves to be from zero to ten. The Bezier curve arc1 is promoted to a Bspline curve.

CROSSEC

PolygonType CROSSEC( PolygonType Object )

This feature is NOT implemented.

CRV2TANS

ListType CRV2TANS( CurveType Crv, NumericType FineNess )

Computes all the bi-tangents of Crv, in the XY plane. That is, all lines that are tangent to Crv at two different locations. Returned is a list of points with X and Y coefficients representinf the parametric locations on Crv of the bi-tangent. FineNess controls the numerical accuracy of the computation. A value of 10 will provide a good start and the larger this number is, the better the accuracy will be. See also CRVPTTAN and CNVXHULL.

Example:

Tans = nil(); Crv2Tns = Crv2Tans( Crv, 10 ); for ( i = 1, 1, sizeof( Crv2Tns ), pt = nth( Crv2Tns, i ): snoc( ceval( Crv, coord( pt, 0 ) ) + ceval( Crv, coord( pt, 1 ) ), Tans ) );

finds the bi-tangents of Crv and convert them to a set of line segments.

CRVLNDST

NumericType CRVLNDST( CurveType Crv, PointType PtOnLine, VectorType LnDir, NumericType IsMinDist, NumericType Epsilon ) or ListType CRVLNDST( CurveType Crv, PointType PtOnLine, VectorType LnDir, NumericType IsMinDist, NumericType Epsilon )

Computes the closest (if IsMinDist is TRUE, farthest if FALSE) point on Curve to the line specified by PtOnLine and LnDir as a point on the line and a line direction. Since this operation is partially numeric, Epsilon is used to set the needed accuracy. It returns the parameter value of the location on Crv closest to the line. If, however, Epsilon is negative, -Epsilon is used instead, and all local extrema in the distance function are returned as a list (both minima and maxima). If the line and the curve intersect, the point of intersection is returned as the minimum.

Example:

Param = CRVLNDST( Crv, linePt, lineVec, TRUE, 0.001 );

finds the closest point on Crv to the line defined by linePt and lineVec.

CRVPTDST

NumericType CRVPTDST( CurveType Crv, PointType Point, NumericType IsMinDist, NumericType Epsilon ) or ListType CRVPTDST( CurveType Crv, PointType Point, NumericType IsMinDist, NumericType Epsilon )

Computes the closest (if IsMinDist is TRUE, farthest if FALSE) point on Crv to Point. Since this operation is partially numeric, Epsilon is used to set the needed accuracy. It returns the parameter value of the location on Crv closest to Point. If, however, Epsilon is negative, -Epsilon is used instead, and all local extrema in the distance function are returned as a list (both minima and maxima).

Example:

Param = CRVPTDST( Crv, Pt, FALSE, 0.0001 );

finds the farthest point on Crv from point Pt.

CRVPTTAN

ListType CRVPTTAN( CurveType Crv, PointType Pt, NumericType FineNess )

Computes all the tangents to Crv that goes through point Pt, all in the XY plane. Returned is a list of points with X and Y coefficients representinf the parametric locations on Crv of the bi-tangent. FineNess controls the numerical accuracy of the computation. A value of 0.01 will provide a good start and the smaller this number is, the better the accuracy will be. See also CRV2TANS and CNVXHULL.

Example:

Tans = nil(); Pt = point( 2, 0, 0 ); CrvPtTns = CrvPtTan( Crv, Pt, 0.01 ); for ( i = 1, 1, sizeof( CrvPtTns ), snoc( ceval( Crv, nth( CrvPtTns, i ) ) + coerce( Pt, e3 ), Tans ) );

finds the tangents of Crv through Pt and convert them to a set of line segments.

CSURFACE

CurveType CSURFACE( SurfaceType Srf, ConstantType Direction, NumericType Param ) or CurveType CSURFACE( TriSrfType Srf, ConstantType Direction, NumericType Param )

Extract an isoparametric curve out of Srf in the specified Direction (ROW or COL (or DEPTH for triangular surface) at the specified parameter value Param. Param must be contained in the parametric domain of Srf in Direction direction. The returned curve is in the surface Srf.

Example:

Crv = CSURFACE( Srf, COL, 0.45 );

extracts an isoparametric curve in the COLumn direction at the parameter value of 0.15 from surface Srf. See also CMESH, COMPOSE.

CTANGENT

VectorType CTANGENT( CurveType Curve, NumericType Param )

Computes the tangent vector to Curve at the parameter value Param. The returned vector has a unit length.

Example:

Tang = CTANGENT( Crv, 0.5 );

computes the tangent vector to Crv at the parameter value of 0.5.

CTLPT

CPt = CTLPT( ConstantType PtType, NumericType Coord1, ... )

Constructs a single control point to be used in the construction of curves and surfaces. Points can have from one to five dimensions, and may be either Euclidean or Projective (rational). Points' type is set via the constants E1 to E5 and P1 to P5. The coordinates of the point are specified in order, weight is first if rational.

Examples:

CPt1 = CTLPT( E3, 0.0, 0.0, 0.0 ); CPt2 = CTLPT( P2, 0.707, 0.707, 0.707 );

constructs an E3 point at the origin and a P2 rational point with a weight of 0.707. The Projective Pi points are specified as CTLPT(Pn, W, W X1, ... , W Xn).

CTRIMSRF

ListType CTRIMSRF( TrimSrfType TSrf, NumericType Parametric )

Extract the trimming curves of a trimmed surface TSrf. If Parametric is not zero, then the trimming curves are extracted as parametric space curves of TSrf. Otherwise, the trimming curves are evaluated into Euclidean space as curves on surface TSrf.

Example:

TrimCrvs = CTRIMSRF( TrimSrf, FALSE );

extracts the trimming curves of TrimSrf as Euclidean curves on TrimSrf.

CYLIN

PolylineType CYLIN( VectorType Center, VectorType Direction, NumericType Radius )

Creates a CYLINder geometric object, defined by Center as center of the base of the CYLINder, Direction as the CYLINder's axis and height, and Radius as the radius of the base of the CYLINder. See RESOLUTION for the accuracy of the CYLINder approximation as a polygonal model. See IRITSTATE's "PrimRatSrfs" and "PrimRatSrfs" state variables.

Example:

Cylinder1 = CYLIN( vector( 0, 0, 0 ), vector( 1, 0, 0 ), 10 );

constructs a cylinder along the X axis from the origin to X = 10.

CZEROS

ListType CZEROS( CurveType Crv, NumericType Epsilon, NumericType Axis )

Computes the zero set of the given Crv in the given axis (1 for X, 2 for Y, 3 for Z). Since this computation is numeric, an Epsilon is also required to specify the desired tolerance. It returns a list of all the parameter values (NumericType) the curve is zero.

Example:

xzeros = CZEROS( cb, 0.001, 1 ); pt_xzeros = nil(); pt = nil(); for ( i = 1, 1, sizeof( xzeros ), pt = ceval( cb, nth( xzeros, i ) ): snoc( pt, pt_xzeros ) ); interact( list( axes, cb, pt_xzeros ), 0 );

Computes the X zero set of curve cb with error tolerance of 0.001. This set is then scanned in a loop and evaluated to the curve's locations, which are then displayed. See also CINFLECT.

EVOLUTE

CurveType EVOLUTE( CurveType Curve ) or SurfaceType EVOLUTE( SurfaceType Curve )

Computes the evolute of a curve or a surface. For curves, the evolute is defined as,

               N(t)

 E(t) = C(t) + ----

               k(t)

 where N(t) is the unit normal of C(t) and k(t) is its curvature.

 E(t) is computed symbolically as the symbolic sum of C(t) and

 N(t) / k(t).

For surfaces, this function computes the mean evulate which is equal to,

                      n(u, v)

 E(u, v) = S(u, v) + ---------

                     2 H(u, v)

 where n(u, v) is the unit normal of S(u, v) and H(u, v) is its mean

 curvature. E(u, v) is computed symbolically.

The result of this symbolic computation is exact (upto machine precision) unlike a similar operations that are only approximated, like the OFFSET or the AOFFSET.

Example:

crv = cbspline( 3, list( ctlpt( E3, -1.0, 0.1, 0.2 ), ctlpt( E3, -0.1, 1.0, 0.1 ), ctlpt( E3, 0.1, 0.1, 1.0 ), ctlpt( E3, 1.0, 0.1, 0.1 ), ctlpt( E3, 0.1, 1.0, 0.2 ) ), list( KV_OPEN ) ); cev = EVOLUTE( Crv );

EXTRUDE

PolygonType EXTRUDE( PolygonType Object, VectorType Dir ) or SurfaceType EXTRUDE( CurveType Object, VectorType Dir )

Creates an extrusion of the given Object. If Object is a PolygonObject, its first polygon is used as the base for the extrusion in Dir direction, and a closed PolygonObject is constructed. If Object is a CurveType, an extrusion surface is constructed instead, which is not a closed object (the two bases of the extrusion are excluded, and the curve may be open by itself).

Direction Dir cannot be coplanar with the polygon plane. The curve may be nonplanar.

Example:

Cross = cbspline( 3, list( ctlpt( E2, -0.018, 0.001 ), ctlpt( E2, 0.018, 0.001 ), ctlpt( E2, 0.019, 0.002 ), ctlpt( E2, 0.018, 0.004 ), ctlpt( E2, -0.018, 0.004 ), ctlpt( E2, -0.019, 0.001 ) ), list( KV_OPEN ) ); Cross = Cross + -Cross * scale( vector( 1, -1, 1 ) ); Napkin = EXTRUDE( Cross * scale( vector( 1.6, 1.6, 1.6 ) ), vector( 0.02, 0.03, 0.2 ) );

constructs a closed cross section Cross by duplicating one half of it in reverse and merging the two sub-curves. Cross is then used as the cross-section for the extrusion operation.

FFCOMPAT

FFCOMPAT( CurveType Crv1, CurveType Crv2 ) or FFCOMPAT( SurfaceType Srf1, SurfaceType Srf2 )

Makes the given two curves or surfaces compatible by making them share the same point type, same curve type, same degree, and the same continuity. Same point type is gained by promoting a lower dimension into a higher one, and non-rational to rational points. Bezier curves are promoted to Bspline curves if necessary, for curve type compatibility. Degree compatibility is achieved by raising the degree of the lower order curve. Continuity is achieve by refining both curves to the space with the same (unioned) knot vector. This function returns nothing and compatibility is made in place.

Example:

FFCOMPAT( Srf1, Srf2 );

See also CMORPH and SMORPH.

FFCTLPTS

ListType FFCTLPTS( CurveType Crv ); or ListType FFCTLPTS( SurfaceType Srf ); or ListType FFCTLPTS( TrimSrfType Srf ); or ListType FFCTLPTS( TrivarType Srf );

Returns all the control points of the given freeform in a single list. See Also FFKNTVEC, FFMSIZE, FFORDER, FFPTTYPE.

Example:

Ctls = FFCTLPTS( Srf1 );

FFEXTREME

CtlPtType FFEXTREME( CurveType Crv, NumericType Minimum ) or CtlPtType FFEXTREME( SurfaceType Srf, NumericType Minimum )

Computes a bound on the extreme values a curves Crv or surface Srf can assume. Returned control points provides a bound on the minimum (maximum) values that can be assumed if Minimum is TRUE (FALSE).

Example:

Bound = FFEXTREME( Srf, false );

Computes a bound on the maximal values Srf can assume.

FFKNTVEC

ListType FFKNTVEC( CurveType Crv ) or ListType FFKNTVEC( SurfaceType Srf ) or ListType FFKNTVEC( TrimSrfType TrimSrf ) or ListType FFKNTVEC( TrivarType Trivar ) or ListType FFKNTVEC( TriSrfType TriSrf )

Returns all the knot vector(s) of the given freeform in a list of knot vector(s). See Also FFCTLPTS, FFMSIZE, FFORDER, FFPTTYPE.

Example:

KVs = FFKNTVEC( Srf1 );

FFMATCH

FFMATCH( CurveType Crv1, CurveType Crv2, NumericType Reduce, NumericType Samples, NumericType ReparamOrder, NumericType Rotate, NumericType NormType )

Computes a reparametrization to Crv2 so it fits Crv1, the best under some prescribed norm, NormType. Currently the following norms are valid for NormType

Value	Description
1	Suitable for ruled and blended curves, for modeling.
	See RULEDSRF.
2	Suitable for metamorphosis of curves. See CMORPH.
3	Distance norm in "walking the dog" notion.
4	Bisector (skeleton) matching norm for two curves.

Whenever negative norms can result (for example, in cases were self intersection cannot be prevented in ruled surface constructions), one can allow negativity with no extra penalty by applaying neative NormType. Use of positive only norms would yield no output at all if no matching with positive weights can be established whereas allowing negative norm values would result in the globally optimal result, but with possibly self intersectiions.

The reparametrization is computed by sampling a fix set of size Samples off both curves, and fitting a Bspline curve of length Reduce as the reparametrization curve. Hence, Reduce must be less than or equal to Samples. The reparametrization curve will have order of ReparamOrder. If Rotate is TRUE or ON, then attempt is made to rotate the reparametrization of the curves. Rotation can be used on closed curves only.

See RULEDSRF and CMORPH for examples.

FFMERGE

CurveType FFMERGE( ListType E1Curves, NumericType PointType )@ or SurfaceType FFMERGE( ListType E1Surfaces, NumericType PointType )

Merges the scalar curves in the list of curves E1Curves or list of surfaces E1Surfaces to one vector curve/surface of point type PointType.

Example:

Srf = FFMERGE( list( SrfW, SrfX, SrfY ), P2 );

merges three scalar surfaces into a single surface with point type P2. See also FFSPLIT, FFPTTYPE.

FFMSIZE

ListType FFMSIZE( CurveType Crv ) or ListType FFMSIZE( SurfaceType Srf ) or ListType FFMSIZE( TrimSrfType TrimSrf ) or ListType FFMSIZE( TrivarType Trivar ) or ListType FFMSIZE( TriSrfType TriSrf )

Returns the size of the control mesh/polygon of the given freeform in a list. See Also FFCTLPTS, FFKNTVEC, FFORDER, FFPTTYPE.

Example:

MSizes = FFMSIZE( Srf1 );

FFORDER

ListType FFORDER( CurveType Crv ) or ListType FFORDER( SurfaceType Srf ) or ListType FFORDER( TrimSrfType TrimSrf ) or ListType FFORDER( TrivarType Trivar ) or ListType FFORDER( TriSrfType TriSrf )

Return all the orders of the given freeform in a single list. See Also FFCTLPTS, FFKNTVEC, FFMSIZE, FFPTTYPE.

Example:

Orders = FFORDER( Srf1 );

FFPOLES

NumericType FFPOLES( CurveType Crv ); or NumericType FFPOLES( SurfaceType Srf ); or NumericType FFPOLES( TrimSrfType TrimSrf ); or NumericType FFPOLES( TrivarType Trivar ); or NumericType FFPOLES( TriSrfType TriSrf );

Returns TRUE in the given freeform has poles, FALSE otherwise. Poles are zeros in the weights of rational functions.

Example:

HasPoles = FFPOLES( Srf1 );

FFPTDIST

ListType FFPTDIST( CurveType Crv, NumericType Euclid, NumericType NumOfPts ) or ListType FFPTDIST( SurfaceType Srf, NumericType Euclid, NumericType NumOfPts )

Computes a uniform point distribution for a Crv or Srf. If Euclid is FALSE the distribution is selected to be uniform in the parametric space, otherwise if TRUE, the distribution is made uniform in the Euclidean space. NumOfPts sets the number of points in the distribution.

The returned list of points prescribes parameters values in the freeforms. For Crv, the returned list is a list of reals, in the parameter space of Crv. For Srf, the returned list is a list of points, whose X and Y coefficients holds the U and V parameters of Srf. See also COVERPT.

Example:

c1 = cbezier( list( ctlpt( E2, -1.0, 0.0 ), ctlpt( E2, -1.0, 0.1 ), ctlpt( E2, -0.9, -0.1 ), ctlpt( E2, 0.9, 0.0 ) ) ); color( c1, magenta ); pts = FFPTDIST( c1, false, 300 ); e2pts = nil(); for ( i = 1, 10, sizeof( pts ), pt = ceval( c1, coord( nth( pts, i ), 0 ) ): snoc( pt, e2pts ) ); interact( list( e2pts, c1 ) ); pts = FFPTDIST( c1, true, 300 ); e2pts = nil(); for ( i = 1, 10, sizeof( pts ), pt = ceval( c1, coord( nth( pts, i ), 0 ) ): snoc( pt, e2pts ) ); interact( list( e2pts, c1 ) );

Computes the distribution of 100 points in curve c1 which has highly nonuniform speed characteristics. Two distributions are computed, one to be uniform in the parametric space and one uniform in the Euclidean space.

FFPTTYPE

NumericType FFPTTYPE( CurveType Crv ) or NumericType FFPTTYPE( SurfaceType Srf ) or NumericType FFPTTYPE( TrimSrfType Srf ) or NumericType FFPTTYPE( TrivarType )

Returns the point type (E2, P4 etc.) of the given freeform. See Also FFCTLPTS, FFKNTVEC, FFMSIZE, FFORDER.

FFSPLIT

ListType FFSPLIT( CurveType Crv ) or ListType FFSPLIT( SurfaceType Srf )

Splits the given curve Crv or surface Srf into its scalar components that are returned as a list of curves/surfaces.

Example:

E1Srfs = FFSPLIT( circle( vector( 0, 0, 0 ), 1 ) );

splits the circle which is a curve in P3 into four scalar curves (W, X, Y, Z) that are returned in a single list. See also FFMERGE, FFPTTYPE.

GBOX

PolygonType GBOX( VectorType Point, VectorType Dx, VectorType Dy, VectorType Dz )

Creates a parallelepiped - Generalized BOX polygonal object, defined by Point as base position, and Dx, Dy, Dz as 3 3D vectors to define the 6 faces of this generalized BOX. The regular BOX object is a special case of GBOX where Dx = vector(Dx, 0, 0), Dy = vector(0, Dy, 0), and Dz = vector(0, 0, Dz).

Dx, Dy, Dz must all be independent in order to create an object with positive volume.

Example:

GB = GBOX( vector( 0.0, -0.35, 0.63 ), vector( 0.5, 0.0, 0.5 ), vector( -0.5, 0.0, 0.5 ), vector( 0.0, 0.7, 0.0 ) );

GETATTR

AnyType GETATTR( AnyType Obj, StringType Name )

Provides a mechanism to fetch an attribute named Name from object Obj.

Example:

attrib( axes, "test", 15 ); a = GETATTR( axes, "test" );

will set the value of a to be 15.

GETLINE

AnyType GETLINE( NumericType RequestedType )

Provides a method to get input from keyboard within functions and or subroutines. RequestedType can be one of NUMERIC_TYPE, POINT_TYPE, VECTOR_TYPE, or PLANE_TYPE in which the entered line will be parsed into one, three, or four numeric values (sperated by either spaces or commas) and the proper object will be created and returned. In any other case, including failure to parse the numeric input, a STRING_TYPE object will be constructed from the entered line.

Example:

Pt = GETLINE( point_type );

to read one point (three numeric values) from stdin.

GPOLYGON

PolygonType GPOLYGON( GeometryTreeType Object, NumericType Normals )

Approximates all Surface(s)/Trimmed surface(s)/Trivariate(s) in Object with polygons using the RESOLUTION and FLAT4PLY variables. The larger the RESOLUTION is, the finer (more polygons) the resulting approximation will be.

FLAT4PLY is a Boolean flag controlling the conversion of an (almost) flat patch into four (TRUE) or two (FALSE) polygons. Normals are computed to polygon vertices using surface normals, so Gouraud or Phong shading can be exploited. It returns a single polygonal object.

If Normals is set, surface normals will be evaluated at the vertices. Otherwise flat shading and constant normals across polygons are assumed.

Example:

Polys = GPOLYGON( list( Srf1, Srf2, Srf3 ), off );

Converts to polygons the three surfaces Srf1, Srf2, and Srf3 with no normals.

GPOLYLINE

PolylineType GPOLYLINE( GeometryTreeType Object, NumericType Optimal )

Converts all Curves(s), (Trimmed) Surface(s), and Trivariate(s) Object into polylines using the RESOLUTION variable. The larger the RESOLUTION is, the finer the resulting approximation will be. It returns a single polyline object.

If Optimal is false, the points are sampled at equally spaced interval in the parametric space. If Optimal true, a better, more expensive computationally, algorithm is used to derive optimal sampling locations as to minimize the maximal distance between the curve and piecewise linear approximation (L infinity norm).

Example:

Polys = GPOLYLINE( list( Srf1, Srf2, Srf3, list( Crv1, Crv2, Crv3 ) ), on );

converts to polylines the three surfaces Srf1, Srf2, and Srf3 and the three curves Crv1, Crv2, and Crv3.

HERMITE

SurfaceType HERMITE( CurveType Bndry1, CurveType Bndry2, CurveType Tan1, CurveType Tan2 ) or CurveType HERMITE( PointType Bndry1, PointType Bndry2, VectorType Tan1, VectorType Tan2 )

Constructs a cubic fit between Bndry1 and Bndry2 so that first derivative continuity constraints, as prescribed by Tan1 at Bndry1 and Tan2 at Bndry2, are preserved.

Returns either a curve or a surface, according to type of input parameters.

Example:

h00 = HERMITE( point( 0, 0, 0 ), point( 1, 1, 0 ), vector( 1, 0, 0 ), vector( 1, 0, 0 ) );

Constructs a curve in the shape of the first basis function of the cubic Hermite basis functions.

INSTANCE

InstanceType INSTANCE( GeometryType Geom );

Creates an instance of the geometry prescribed by Geom. The instance is created with a unit transformation matrix that can be modified using a product with a matrix object.

The use of instances is advantageous where the same geometry is to be displayed/processed in several different locations in space. A modification of the original geometry Geom will affect all instances that reference it. The reference is by the original object's name. The original object can be a single object or a whole hierarchy of objects.

Example:

Tea1 = INSTANCE( "Teapot" ) * trans( vector( 10, 0, 0 ) ); Tea2 = INSTANCE( "Teapot" ) * trans( vector( 20, 0, 0 ) ); Tea3 = INSTANCE( "Teapot" ) * trans( vector( 30, 0, 0 ) ); viewobj( list( Teapot, Tea1, Tea2, Tea3 ) );

will display four teapots 10 units apart in X.

As an instance is affected by an application of a transformation to it, if both the original object and the instance are in the same hierarchy, the transformation of the hierarchy will affect the instance twice:

chair = chair(); # Creates a chair object using chair function chairInst = INSTANCE( "chair" ) * tx( 1 ); Hier = list( chair, chairInst ) * ty( 1 );

will, in Hier, translate chair to (0, 1, 0) but will translate chairInst to (1, 2, 0) or translate twice in y (once chair and once chairInst).

LOFFSET

CurveType LOFFSET( CurveType Crv, NumericType OffsetDistance, NumericType NumOfSamples, NumericType NumOfDOF, NumericType Order )

Approximate an offset of OffsetDistance by sampling NumOfSamples samples along the offset curve and least square fitting them using a Bspline curve of order Order and NumOfDOF control points.

Example:

OffCrv1 = LOFFSET( Crv, -0.4, 100, 10, 4 );

See also OFFSET, AOFFSET, and MOFFSET.

MERGPOLY

PolygonType MERGEPOLY( ListType PolyList )

Merges a set of polygonal objects in PolyList list to a single polygonal object. All elements in ObjectList must be of PolygonType type. This function performs the same operation as the overloaded ^ operator would, but might be more convenient to use under some circumstances.

Example:

Vrtx1 = vector( -3, -2, -1 ); Vrtx2 = vector( 3, -2, -1 ); Vrtx3 = vector( 3, 2, -1 ); Vrtx4 = vector( -3, 2, -1 ); Poly1 = poly( list( Vrtx1, Vrtx2, Vrtx3, Vrtx4 ), false ); Vrtx1 = vector( -3, 2, 1 ); Vrtx2 = vector( 3, 2, 1 ); Vrtx3 = vector( 3, -2, 1 ); Vrtx4 = vector( -3, -2, 1 ); Poly2 = poly( list( Vrtx1, Vrtx2, Vrtx3, Vrtx4 ), false ); Vrtx1 = vector( -3, -2, 1 ); Vrtx2 = vector( 3, -2, 1 ); Vrtx3 = vector( 3, -2, -1 ); Vrtx4 = vector( -3, -2, -1 ); Poly3 = poly( list( Vrtx1, Vrtx2, Vrtx3, Vrtx4 ), false ); PolyObj = MERGEPOLY( list( Poly1, Poly2, Poly3 ) );

MOFFSET

CurveType MOFFSET( CurveType Crv, NumericType OffsetDistance, NumericType AngularError )

Computes an offset of OffsetDistance with globally bounded error (controlled by AngularError). The smaller AngularError is, the better the approximation to the offset. The bounded error is achieved by adaptive refinement of the Crv. The offset is computed via matching of the tangent fields of the given curve Crv and an arc spanning the same angular domain. Further, AngularError measures the angular deviation allowed between the two tangent fields.

Example:

OffCrv1 = MOFFSET( Crv, -0.4, 10 ); OffCrv2 = MOFFSET( Crv, -0.4, 5 );

computes an offset approximation to Crv with OffsetDistance of -0.4 and AngularError of 10 and 5 degrees, respectively. See also OFFSET, AOFFSET, LOFFSET, and FFMATCH.

MOMENT

PointType MOMENT( CurveType Crv, 0 ); or VectorType MOMENT( CurveType Crv, 1 );

Approximates the zero and first moment of curve Crv.

Example:

a = circle( vector( 0, 0, 0 ), 1 ); a = cregion( a, 0, 1 ); p = moment( a, 0 ); v = moment( a, 1 ); view(list(a, p, v), on); a = cregion( a, 0, 1 ) * rz( 45 ); p = moment( a, 0 ); v = moment( a, 1 ); view(list(a, p, v), on);

computes and displays the zero and first moment of a quarter of a circle in two orientations.

MRCHCUBE

PolygonType MRCHCUBE( ListType VolumeSpec, PointType CubeDim, NumericType SkipFactor, NumericType IsoVal )

Applies (a variation of) the marching cubes algorithm (see W. E.Lorensen and H. E.Cline. "Marching Cubes: A High Resolution 3D Surface Construction Algorithm." Computer Graphics (SIGGRAPH '87 Proceedings), Vol. 21, No. 4, pp 163-169, July 1987.) to the given volumetric data set or trivariate. VolumeSpec can be a list of three or five objects as follows:

3	a triplet of the form
	(TrivarType TV, NumericType Axis, NumericType TVNormal)
5	a 5-tuple of the form
	(StringType FileName, NumericType DataType, NumericType Width,
	NumericType Height, NumericType Depth )

In the first case, the trivariate TV is iso surface contoured at level IsoVal along the prescribed Axis (Note a trivariate need not be a scalar function, while Marching Cubes assumes a scalar function). If TVNormals is not zero, much more accurate normals are derived using the trivariate function which is also slower. Otherwise, first order differencing on the cubes is employed for normal estimation.

In the second case, the volume file prescribed by FileName is loaded and iso surface contoured. The file is assumed to hold Width * Height * Depth (Width first, Depth order last) scalar numeric values of type DataType:

1	Regular ascii (separated by while spaces)
2	Two bytes short integer.
3	Four bytes long integer.
4	One byte (char) integer.
5	Four bytes float.
6	Eight bytes double.

Beware of the little vs big Endian problem! We assume here you read in the volume in the same machine type this file was writeen with.

CubeDim allows the user to prescribe the real cell size (not necessarily cubical. SkipFactor allow the skipping ofdata in large data sets. SkipFactor = 1 skips nothing. SkipFactor = 2, skips every other scalar value, reducing in half all dimensions, etc. Last but not least, IsoVal sets the iso surface level.

See also COVERISO, TVLOAD, and TMORPH.

Example:

IsoSrf = MRCHCUBE( list( ThreeCyls, 1, TRUE ), point( 1, 1, 1 ), 1, 0.12 );

Iso surface contour the X axis of trivariate ThreeCyls and use the trivariate to get better normal's estimations. Cell size is unit cube like, no skipping of data and iso surface level is 0.12.

NIL

ListType NIL()

Creates an empty list so data can be accumulated in it. See CINFLECT or CZEROS for examples. See also LIST and SNOC.

OFFSET

PolygonType OFFSET( PolygonType Poly, NumericType OffsetDistance, NumericType Tolerance, NumericType BezInterp ) or CurveType OFFSET( CurveType Crv, NumericType OffsetDistance, NumericType Tolerance, NumericType BezInterp ) or SurfaceType OFFSET( SurfaceType Srf, NumericType OffsetDistance, NumericType Tolerance, NumericType BezInterp ) or TrimSrfType OFFSET( TrimSrfType TrimSrf, NumericType OffsetDistance, NumericType Tolerance, NumericType BezInterp )

Offsets Poly, Crv, Srf or a TrimSrf, by translating all the (control) points in the direction of the normal of the poly/curve or of the (trimmed) surface by an OffsetDistance amount. For a Poly object, result is an offset of the original polygon/line in the XY plane and is exact. Hence, Tolerance and BezInterp parameters are ignored. Result is a mitter joint based offset. Otherwise, each control point has a node parameter value associated with it, which is used to compute the normal. The returned curve or surface only approximates the real offset. If the resulting approximation does not satisfy the accuracy required by Tolerance, Crv or Srf or TrimSrf is subdivided and an offset approximation fit is computed to the two halfs. For curves, one can request a Bezier interpolation scheme in the offset approximation by setting BezInterp. The BezInterp is not supported yet for (trimmed) surfaces. Negative OffsetDistance denotes offset in the reversed direction of the normal.

Example:

OffCrv = OFFSET( Crv, -0.4, 0.1, off );

offsets Crv by the amount of -0.4 in the reversed normal direction, Tolerance of 0.1 and no Bezier interpolation. See also AOFFSET and LOFFSET.

ORTHOTOMC

CurveType ORTHOTOMC( CurveType Crv, PointType Pt, NumericType K ) or, SurfaceType ORTHOTOMC( SurfaceType Srf, PointType Pt, NumericType K )

Computes the K-orthotomic of freeform curves and surfaces. See Fundamentals of Computer Aided Geometric Design, by J. Hoschek and and D. Lasser.

Example:

pt = point( 0, 0.35, 0 ); crv = cbezier( list( ctlpt( E2, -0.8, -0.6 ), ctlpt( E2, -0.3, -0.2 ), ctlpt( E2, 0.0, 0.0 ), ctlpt( E2, 0.8, -0.6 ) ) ); Orth = ORTHOTOMC( crv, pt, 2 ); interact( list( Orth, crv, pt ) * tx( 0.5 ) ) );

computes the orthotomic of a cubic Bezier curve that has an inflection point. Note inflection points are reduced to cusps in the orthotomic result.

PCIRCLE

CurveType PCIRCLE( VectorType Center, NumericType Radius )

Same as CIRCLE but approximates the circle as a polynomial curve. See CIRCLE.

PDECIMATE

PolygonType PDECIMATE( PolygonType Obj, NumericType Dist, NumericType NumPasses, NumericType DecimRatio, NumericType MinAspRatio )

Given a polygonal model, Obj, decimate and merge polygons, effectively reducing the size of the data subject to a maximal deviation distance Dist, number of reduction passes NumPasses, maximal number of DecimRatio of reductions per vertex, and aspect ratio of the triangulation, MinAspRatio.

Example:

gcross = cbspline( 3, list( ctlpt( E3, 0.3, 0.0, 0.0 ), ctlpt( E3, 0.1, 0.0, 0.1 ), ctlpt( E3, 0.1, 0.0, 0.4 ), ctlpt( E3, 0.5, 0.0, 0.5 ), ctlpt( E3, 0.6, 0.0, 0.8 ) ), list( KV_OPEN ) ); resolution = 30; glass = surfprev( gcross ); pglass = gpolygon( glass, false ); dglass = PDECIMATE( pglass, 0.005, 3, 1, 0 );

creates a surface of a glass, approximate it with polygons and then decimates the latter.

PDOMAIN

ListType PDOMAIN( CurveType Crv ) or ListType PDOMAIN( SurfaceType Srf ) or ListType PDOMAIN( TriSrfType Srf ) or ListType PDOMAIN( TrimSrfType TrimSrf ) or ListType PDOMAIN( TrivarType TV )

Returns the parametric domain of the curve (TMin, TMax) or of a (trimmed) surface (UMin, UMax, VMin, VMax) or of a triangular surface or trivariate function (UMin, UMax, VMin, VMax, WMin, WMax) as a list object.

Example:

circ_domain = PDOMAIN( circle( vector( 0.0, 0.0, 0.0 ), 1.0 ) );

PLN3PTS

PlaneType PLN3PTS( PointType Pt1, PointType Pt2, PointType Pt3 )

Computes a plane out of three points.

Example:

Pl1 = PLN3PTS( point( 0, 0, 0 ), point( 0, 1, 0 ), point( 1, 0, 0 ) );

POLY

PolygonType POLY( ListType VrtxList, NumericType IsPolyline )

Creates a single polygon/polyline (and therefore open) object, defined by the vertices in VrtxList (see LIST). All elements in VrtxList must be of VectorType type. If IsPolyline, a polyline is created, otherwise a polygon.

Example:

V1 = vector( 0.0, 0.0, 0.0 ); V2 = vector( 0.3, 0.0, 0.0 ); V3 = vector( 0.3, 0.0, 0.1 ); V4 = vector( 0.2, 0.0, 0.1 ); V5 = vector( 0.2, 0.0, 0.5 ); V6 = vector( 0.3, 0.0, 0.5 ); V7 = vector( 0.3, 0.0, 0.6 ); V8 = vector( 0.0, 0.0, 0.6 ); V9 = vector( 0.0, 0.0, 0.5 ); V10 = vector( 0.1, 0.0, 0.5 ); V11 = vector( 0.1, 0.0, 0.1 ); V12 = vector( 0.0, 0.0, 0.1 ); I = POLY( list( V1, V2, V3, V4, V5, V6, V7, V8, V9, V10, V11, V12 ), FALSE );

constructs an object with a single polygon in the shape of the letter I.

PRISA

ListType PRISA( SurfaceType Srfs, NumericType SamplesPerCurve, NumericType Epsilon, ConstantType Dir, VectorType Space )

Computes a layout (prisa) of the given surface(s) Srfs, and returns a list of surface objects representing the layout. The surface is approximated to within Epsilon in direction Dir into a set of ruled surfaces and then developable surfaces that are laid out flat onto the XY plane. If Epsilon is negative, the piecewise ruled surface approximation in 3-space is returned. SamplesPerCurve controls the piecewise linear approximation of the boundary of the ruled/developable surfaces. Space is a vector whose X component controls the space between the different surfaces' layout, and whose Y component controls the space between different layout pieces.

Example:

cross = cbspline( 3, list( ctlpt( E3, 0.7, 0.0, 0. ), ctlpt( E3, 0.7, 0.0, 0.06 ), ctlpt( E3, 0.1, 0.0, 0.1 ), ctlpt( E3, 0.1, 0.0, 0.6 ), ctlpt( E3, 0.6, 0.0, 0.6 ), ctlpt( E3, 0.8, 0.0, 0.8 ), ctlpt( E3, 0.8, 0.0, 1.4 ), ctlpt( E3, 0.6, 0.0, 1.6 ) ), list( KV_OPEN ) ); wglass = surfrev( cross ); wgl_ruled = PRISA( wglass, 6, -0.1, COL, vector( 0, 0.25, 0.0 ) ); wgl_prisa = PRISA( wglass, 6, 0.1, COL, vector( 0, 0.25, 0.0 ) );

Computes a layout of a wine glass in wgl_prisa and a three-dimensional ruled surface approximation of wglass in wgl_ruled.

PT3BARY

VectorType PT3BARY( PointType Pt1, PointType Pt2, PointType Pt3, PointType InteriorPt )

Computes the barycentric coordinates of InterPt with respect to the triangle defined by Pt1, Pt2, Pt3. Returned is a vector of three coefficents, which are the weights of the three points of the triangle. InteriorPt is assumed to be in the triangle.

Example:

Coeffs = PT3BARY( point( 0, 0, 0 ), point( 1, 0, 0 ), point( 0, 1, 0 ), point( 0.25, 0.25, 0.0 ) );

PTHMSPR

ListType PTHMSPR( NumericType Size )

Computes a fairly uniform distribution of points on a hemisphere. Size hints on the distance between adjacent placed points.

Example:

Pts = PTHMSPR( 0.1 );

PTLNPLN

VectorType PTLNPLN( PointType LineOrig, VectorType LineRay, PlaneType Plane )

Computes the point of intersection of given line LineOrig, LineRay with plane Plane.

Example:

InterPt = PtLnPln( point( 1, 0, 1 ), vector( 1, 1, 1 ), Plane( 0, 0, 1, 0 ) );

PTPTLN

VectorType PTPTLN( PointType Point, PointType LineOrig, VectorType LineRay )

Computes the point on line LineOrig, LineRay that is closest to point Point. See also DSTPTLN

Example:

ClosestPt = PTPTLN( point( 0, 0, 0 ), point( 1, 1, 0 ), vector( 1, 1, 1 ) );

PTSLNLN

ListType PTSLNLN( PointType Line1Orig, VectorType Line1Ray, PointType Line2Orig, VectorType Line2Ray )

Computes the closest two points on the two lines defined by point LineiOrig and ray LineiRay. See also DSTLNLN. Returned is a list object with the two points.

Example:

ClosestPts = PtsLnLn( point( 1, 0, 0 ), vector( 0, 1, 0 ), point( 0, 1, 0 ), vector( 1, 0, 0 ) );

RULEDSRF

SurfaceType RULEDSRF( CurveType Crv1, CurveType Crv2 ) or PolygonType RULEDSRF( PolygonType Poly1, PolygonType Poly22 )

Constructs a ruled surface between the two curves Crv1 and Crv2 or two polylines Poly1 and Poly2. The curves do not have to have the same order or type, and will be promoted to their least common denominator. The polys must have the same number of points and both must be either polygons or polylines.

Example:

c1 = cbspline( 3, list( ctlpt(E3, 1.7, 0.0 , 0 ), ctlpt(E3, 0.7, 0.7 , 0 ), ctlpt(E3, 1.7, 0.3 , 0 ), ctlpt(E3, 1.5, 0.8 , 0 ), ctlpt(E3, 1.6, 1.0 , 0 ) ), list( KV_OPEN ) ); c2 = cbspline( 3, list( ctlpt(E3, 0.7, 0.0 , 0 ), ctlpt(E3,-0.7, 0.2 , 0 ), ctlpt(E3, 0.7, 0.5 , 0 ), ctlpt(E3,-0.7, 0.7 , 0 ), ctlpt(E3, 0.7, 1.0 , 0 ) ) , list( KV_OPEN ) ); srf1 = RULEDSRF( c1, c2 ); interact( list( c1, c2, srf1 ), on ); c2a = ffmatch( c1, c2, 50, 100, 2, false, 1 ); srf2 = RULEDSRF( c1, c2a ); interact( list( c1, c2, srf2 ), on );

Constructs a planar ruled surface between two curves, c1 and c2. The naive construction causes self intersection, but by employing FFMATCH the self intersection can be resloved.

See also FFMATCH.

SBEZIER

SurfaceType SBEZIER( ListType CtlMesh )

Creates a Bezier surface using the provided control mesh. CtlMesh is a list of rows, each of which is a list of control points. All control points must be of type (E1-E5, P1-P5), or regular PointType defining the surface's control polygon. Surface's point type will be of a space which is the union of the spaces of all points.

Example:

Srf = SBEZIER( list ( list( ctlpt( E3, 0.0, 0.0, 1.0 ), ctlpt( E3, 0.0, 1.0, 0.0 ), ctlpt( E3, 0.0, 2.0, 1.0 ) ), list( ctlpt( E3, 1.0, 0.0, 0.0 ), ctlpt( E3, 1.0, 1.0, 2.0 ), ctlpt( E3, 1.0, 2.0, 0.0 ) ), list( ctlpt( E3, 2.0, 0.0, 2.0 ), ctlpt( E3, 2.0, 1.0, 0.0 ), ctlpt( E3, 2.0, 2.0, 2.0 ) ), list( ctlpt( E3, 3.0, 0.0, 0.0 ), ctlpt( E3, 3.0, 1.0, 2.0 ), ctlpt( E3, 3.0, 2.0, 0.0 ) ), list( ctlpt( E3, 4.0, 0.0, 1.0 ), ctlpt( E3, 4.0, 1.0, 0.0 ), ctlpt( E3, 4.0, 2.0, 1.0 ) ) ) );

SBISECTOR

SurfaceType SBISECTOR( SurfaceType Srf, PointType Pt )

Computes the bisector surface of a given surface a a point. See also CBISECTOR.

Example:

s = ruledSrf( ctlpt( E3, -1.0, -1.0, 0.0 ) + ctlpt( E3, 1.0, -1.0, 0.0 ), ctlpt( E3, -1.0, 1.0, 0.0 ) + ctlpt( E3, 1.0, 1.0, 0.0 ) ); pt = point( 0.0, 0.0, 1.0 ); bisect = SBISECTOR( s, pt ); interact( list( s, pt, bisect ) );

computes the bisector surface of a plane and a point.

SBSPLINE

SurfaceType SBSPLINE( NumericType UOrder, NumericType VOrder, ListType CtlMesh, ListType KnotVectors )

Creates a Bspline surface from the provided UOrder and VOrder orders, the control mesh CtlMesh, and the two knot vectors KnotVectors. CtlMesh is a list of rows, each of which is a list of control points. All control points must be of point type (E1-E5, P1-P5), or regular PointType defining the surface's control mesh. Surface's point type will be of a space which is the union of the spaces of all points. KnotVectors is a list of two knot vectors. Each knot vector is a list of NumericType knots of length #CtlPtList plus the Order. If, however, the length of the knot vector is equal to #CtlPtList + Order + Order - 1} the curve is assumed periodic. The knot vector may also be a list of a single constant KV_OPEN or KV_FLOAT or KV_PERIODIC, in which a uniform knot vector with the appropriate length and with open, floating or periodic end condition will be constructed automatically.

Example:

Mesh = list ( list( ctlpt( E3, 0.0, 0.0, 1.0 ), ctlpt( E3, 0.0, 1.0, 0.0 ), ctlpt( E3, 0.0, 2.0, 1.0 ) ), list( ctlpt( E3, 1.0, 0.0, 0.0 ), ctlpt( E3, 1.0, 1.0, 2.0 ), ctlpt( E3, 1.0, 2.0, 0.0 ) ), list( ctlpt( E3, 2.0, 0.0, 2.0 ), ctlpt( E3, 2.0, 1.0, 0.0 ), ctlpt( E3, 2.0, 2.0, 2.0 ) ), list( ctlpt( E3, 3.0, 0.0, 0.0 ), ctlpt( E3, 3.0, 1.0, 2.0 ), ctlpt( E3, 3.0, 2.0, 0.0 ) ), list( ctlpt( E3, 4.0, 0.0, 1.0 ), ctlpt( E3, 4.0, 1.0, 0.0 ), ctlpt( E3, 4.0, 2.0, 1.0 ) ) ); Srf = SBSPLINE( 3, 3, Mesh, list( list( KV_OPEN ), list( 3, 3, 3, 4, 5, 6, 6, 6 ) ) );

constructs a bi-quadratic Bspline surface with its first knot vector having uniform knot spacing with open end conditions.

SCRVTR

SurfaceType SCRVTR( SurfaceType Srf, ConstType PtType, ConstType Dir )

Symbolically computes the extreme curvature bound on Srf. If Dir is either ROW or COL, then the normal curvature square of Srf in Dir is computed symbolically and returned. Otherwise, a upper bound on the sum of the squares of the two principal curvatures is symbolically computed and returned.

Returned value is a surface that can be evaluated to the curvature bound, given a UV location. The returned surface value is a scalar field of point type P1 (scalar rational). However, if PtType is one of E1, P1, E3, P3 the returned surface is coerced to this given type. If the types are one of E3, P3, then the Y and Z axes are set to be equivalent to the U and V parametric domains.

   This function computes the square of the normal curvature scalar

 field for surfaces as (in the U parametric direction, same for V),

d S

              < n , --- >

    u               du

   k (u, v) = ------------

                 dS   ds

               < -- , -- >

                 du   du

 and computes the sum of the squares of the principal curvatures as,

  2    2  ( g11 l22 + g22 l11 - 2 g12 l12 )^2 - 2 |G| |L|

 k  + k = -----------------------------------------------

  1    2                   |G|^2 ||n||^2

See also CCRVTR.

Example:

cross = cbspline( 3, list( ctlpt( E2, 0.0, 0.0 ), ctlpt( E2, 0.8, 0.0 ), ctlpt( E2, 0.8, 0.2 ), ctlpt( E2, 0.07, 1.4 ), ctlpt( E2, -0.07, 1.4 ), ctlpt( E2, -0.8, 0.2 ), ctlpt( E2, -0.8, 0.0 ), ctlpt( E2, 0.0, 0.0 ) ), list( KV_OPEN ) ); cross = coerce( cross, e3 ); s = sFromCrvs( list( cross, cross * trans( vector( 0.5, 0, 1 ) ), cross * trans( vector( 0, 0, 2 ) ) ), 3 ); view( list( s, axes ), on ); UCrvtrZXY = scrvtr( s, E3, row ); VCrvtrZXY = scrvtr( s, E3, col ); UCrvtrXYZ = UCrvtrZXY * rotx( -90 ) * roty( -90 ) * scale( vector( 1, 1, 0.001 ) ); VCrvtrXYZ = VCrvtrZXY * rotx( -90 ) * roty( -90 ) * scale( vector( 1, 1, 10 ) ); color( UCrvtrXYZ, red ); color( VCrvtrXYZ, magenta ); view( list( UCrvtrXYZ, VCrvtrXYZ ), off ); CrvtrZXY = scrvtr( s, E3, off ); CrvtrXYZ = CrvtrZXY * rotx( -90 ) * roty( -90 ) * scale( vector( 1, 1, 0.001 ) ); color( CrvtrXYZ, green ); view( CrvtrXYZ, off );

Computes the sqaure of the normal curvature in the U and V direction, flips its scalar value from X to Z using rotations and scale the fields to reasonable values and display them. Then, display a total bound on the normal curvature as well.

Due to the large degree of the resulting fields be warned that rational surfaces will compute into large degree curvature bound fields. See also IRITSTATE("InterpProd", FALSE); for faster symbolic computation.

SDERIVE

SurfaceType SDERIVE( SurfaceType Srf, NumericType Dir )

Returns a vector field surface representing the differentiated surface in the given direction (ROW or COL). Evaluation of the returned surface at a given parameter value will return a vector tangent to Srf in Dir at that parameter value.

DuSrf = SDERIVE( Srf, ROW ); DvSrf = SDERIVE( Srf, COL ); Normal = coerce( seval( DuSrf, 0.5, 0.5 ), VECTOR_TYPE ) ^ coerce( seval( DvSrf, 0.5, 0.5 ), VECTOR_TYPE );

computes the two partial derivatives of the surface Srf and computes its normal as their cross product, at the parametric location (0.5, 0.5).

SDIVIDE

SurfaceType SDIVIDE( SurfaceType Srf, ConstantType Direction, NumericType Param ) or TrimSrfType SDIVIDE( TrimSrfType Srf, ConstantType Direction, NumericType Param )

Subdivides a (possibly trimmed) surface into two at the specified parameter value Param in the specified Direction (ROW or COL). Srf can be either a Bspline surface in which Param must be conatined in the parametric domain of the surface, or a Bezier surface in which Param} must be in the range of zero to one.

It returns a list of upto two sub-surfaces. The individual surfaces may be extracted from the list using the NTH command. If Srf is a trimmed surface, it can be the case that one of the two subdivided surfaces is completely trimmed out, and hence only one surface will be returned.

Example:

SrfLst = SDIVIDE( Srf, ROW, 0.5 ); Srf1 = nth( SrfLst, 1 ); Srf2 = nth( SrfLst, 2 );

Subdivides Srf at the parameter value of 0.5 in the ROW direction.

SEDITPT

SurfaceType SEDITPT( SurfaceType Srf, CtlPtType CPt, NumericType UIndex, NumericType VIndex )

Provides a simple mechanism to manually modify a single control point number UIndex and VIndex (base count is 0) in the control mesh of Srf by substituting CtlPt instead. CtlPt must have the same point type as the control points of Srf. Original surface Srf is not modified.

Example:

CPt = ctlpt( E3, 1, 2, 3 ); NewSrf = SEDITPT( Srf, CPt, 0, 0 );

Constructs a NewSrf with the first control point of Srf being CPt.

SEVAL

CtlPtType SEVAL( SurfaceType Srf, NumericType UParam, NumericType VParam ) or CtlPtType SEVAL( TrimSrfType Srf, NumericType UParam, NumericType VParam )

Evaluates the provided (possibly trimmed) surface Srf at the given UParam and VParam parameters. Both UParam and VParam should be contained in the surface parametric domain if Srf is a Bspline surface, or between zero and one if Srf is a Bezier surface. The returned control point has the same type as the control points of Srf.

Example:

CPt = SEVAL( Srf, 0.25, 0.22 );

Evaluates Srf at the parameter values of (0.25, 0.22).

SFOCAL

SurfaceType SFOCAL( SurfaceType Srf, NumericType Dir )

Evaluates the focal surface field of surface Srf using the normal curvature in the isoparametric direction as given by Dir (either ROW or COL). Note this function is not using the principal curvatures as is generaly the case for focal surfaces.

Example:

gcross = cbspline( 3, list( ctlpt( E3, 0.3, 0.0, 0.0 ), ctlpt( E3, 0.1, 0.0, 0.1 ), ctlpt( E3, 0.1, 0.0, 0.4 ), ctlpt( E3, 0.5, 0.0, 0.5 ), ctlpt( E3, 0.6, 0.0, 0.8 ) ), list( KV_OPEN ) ); glass = surfprev( gcross ); color( glass, red ); gfocal = SFOCAL(glass, col);

Evaluates the focal surface using the COL isoparametric direction's normal curvature of the glass surface.

SFROMCRVS

SurfaceType SFROMCRVS( ListType CrvList, NumericType OtherOrder )

Constructs a surface by substituting the curves in CrvList as rows in a control mesh of a surface. Curves in CrvList are made compatible by promoting Bezier curves to Bsplines if necessary, and raising degree and refining as required before substituting the control polygons of the curves as rows in the mesh. The other direction order is set by OtherOrder, which cannot be larger than the number of curves.

The surface interpolates the first and last curves only.

Example:

Crv1 = cbspline( 3, list( ctlpt( E3, 0.0, 0.0, 0.0 ), ctlpt( E3, 1.0, 0.0, 0.0 ), ctlpt( E3, 1.0, 1.0, 0.0 ) ), list( KV_OPEN ) ); Crv2 = Crv1 * trans( vector( 0.0, 0.0, 1.0 ) ); Crv3 = Crv2 * trans( vector( 0.0, 1.0, 0.0 ) ); Srf = SFROMCRVS( list( Crv1, Crv2, Crv3 ), 3 );

SGAUSS

SurfaceType SGAUSS( SurfaceType Srf )

Evaluates the Gaussian curvature field of surface Srf.

Example:

Srf1 = hermite( cbezier( list( ctlpt( E3, 0.0, 0.0, 0.0 ), ctlpt( E3, 0.5, 0.2, 0.0 ), ctlpt( E3, 1.0, 0.0, 0.0 ) ) ), cbezier( list( ctlpt( E3, 0.0, 1.0, 0.0 ), ctlpt( E3, 0.5, 0.8, 0.0 ), ctlpt( E3, 1.0, 1.0, 0.5 ) ) ), cbezier( list( ctlpt( E3, 0.0, 2.0, 0.0 ), ctlpt( E3, 0.0, 2.0, 0.0 ), ctlpt( E3, 0.0, 2.0, 0.0 ) ) ), cbezier( list( ctlpt( E3, 0.0, 2.0, 0.0 ), ctlpt( E3, 0.0, 2.0, 0.0 ), ctlpt( E3, 0.0, 2.0, 0.0 ) ) ) ); SGauss = SGAUSS( Srf1 );

Evaluates the Gaussian curvaure of Srf1.

SINTERP

SurfaceType SINTERP( ListType PtList, NumericType UOrder, NumericType VOrder, NumericType USize, NumericType VSize, ConstantType Param)

Computes a Bspline polynomial surface that interpolates or approximates the rectangular grid or scattered set of points in PtList. The Bspline surface will have orders UOrder and VOrder and mesh of size USize by VSize control points. If the data is on a grid, the knots will be spaced according to Param which can be one of PARAM_UNIFORM, PARAM_CHORD or PARAM_CENTRIP. The former prescribed a uniform knot sequence and the latters specified a knot spacing according to the chord length and a square root of the chord length. Currently only PARAM_UNIFORM is supported. For scattered point set, the Param parameter is ignored. PtList is a list of list of points for girdded data where all lists should carry the same amount of points in them, defining a rectangular grid. For scattered data, PtList is a linear list of points. All points in PtList must be of type (E1-E5, P1-P5) control point, or regular PointType. If USize and VSize are equal to the number of points in the grid data set of PtList, the resulting surface will interpolate the data set. Otherwise, if USize or VSize is less than the number of points in the grid of PtList, the point data set will be least square approximated. In no time can USize or VSize be larger that the number of points in PtList or lower than UOrder and VOrder, respectively. If USize or VSize are zero, the grid size is used, forcing an interpolation of the data set. If PtList contains a linear list of points, these points are treated as scattered. Each scattered point is assumed to be holding the parameteric location to interpolate at as its first two coefficients. The other coefficients are the interpolation values. In other words, to interpolated scattered data of type E3, E5 control points in a linear list must be provided in (u, v, x, y, z) format. Scattered data is interpolated over a unit square (0 to 1) parameteric domain in both u and v.

All interior knots will be distinct preserving maximal continuity. The resulting Bspline surface will have open end conditions.

Example:

pl = nil(); pll = nil(); for ( x = -5, 1, 5, pl = nil(): for ( y = -5, 1, 5, snoc( point( x, y, sin( x * Pi / 2 ) * cos( y * Pi / 2 ) ), pl ) ): snoc( pl, pll ) ); s1 = sinterp( pll, 3, 3, 8, 8, PARAM_UNIFORM ); s2 = sinterp( pll, 3, 3, 11, 11, PARAM_UNIFORM );

Samples an explicit surface sin(x) * cos(y) at a grid of 11 by 11 points, least square fit with a grid of size of 8 by 8 surface s1, and interpolate surface s2 using this data set.

SMEANSQR

SurfaceType SMEANSQR( SurfaceType Srf )

Evaluates the square of the mean curvature field of surface Srf.

Example:

Srf1 = hermite( cbezier( list( ctlpt( E3, 0.0, 0.0, 0.0 ), ctlpt( E3, 0.5, 0.2, 0.0 ), ctlpt( E3, 1.0, 0.0, 0.0 ) ) ), cbezier( list( ctlpt( E3, 0.0, 1.0, 0.0 ), ctlpt( E3, 0.5, 0.8, 0.0 ), ctlpt( E3, 1.0, 1.0, 0.5 ) ) ), cbezier( list( ctlpt( E3, 0.0, 2.0, 0.0 ), ctlpt( E3, 0.0, 2.0, 0.0 ), ctlpt( E3, 0.0, 2.0, 0.0 ) ) ), cbezier( list( ctlpt( E3, 0.0, 2.0, 0.0 ), ctlpt( E3, 0.0, 2.0, 0.0 ), ctlpt( E3, 0.0, 2.0, 0.0 ) ) ) ); SMean = SMEANSQR( Srf1 );

Evaluates the square of the mean curvaure of Srf1.

SMERGE

SurfaceType SMERGE( SurfaceType Srf1, SurfaceType Srf2, NumericType Dir, NumericType SameEdge )

Merges two surfaces along the requested direction (ROW or COL). If SameEdge is non-zero (ON or TRUE), then the common edge is assumed to be identical and copied only once. Otherwise (OFF or FALSE), a ruled surface is constructed between the two surfaces along the (not) common edge.

Example:

MergedSrf = SMERGE( Srf1, Srf2, ROW, TRUE );

SMORPH

SurfaceType SMORPH( SurfaceType Srf1, SurfaceType Srf2, NumericType Blend )

Creates a new surface which is a convex blend of the two given surfaces. The two given surfaces must be compatible (see FFCOMPAT) before this blend is invoked. Very useful if a sequence that "morphs" one surface to another is to be created.

Example:

for ( i = 0.0, 1.0, 11.0, Msrf = SMORPH( Srf1, Srf2, i / 11.0 ): color( Msrf, white ): attrib( Msrf, "rgb", "255,255,255" ): attrib( Msrf, "reflect", "0.7" ): save( "morp1-" + i, Msrf ) );

creates a sequence of 12 surfaces, morphed from Srf1 to Srf2 and saves them in the files "morph-0.dat" to "morph-11.dat". See also CMORPH and TMORPH.

SNORMAL

VectorType SNORMAL( SurfaceType Srf, NumericType UParam, NumericType VParam ) or VectorType SNORMAL( TrimSrfType Srf, NumericType UParam, NumericType VParam )

Computes the normal vector to (possibly trimmed) surface Srf at the parameter values UParam and VParam. The returned vector has a unit length.

Example:

Normal = SNORMAL( Srf, 0.5, 0.5 );

computes the normal to Srf at the parameter values (0.5, 0.5). See also SNRMLSRF.

SNRMLSRF

SurfaceType SNRMLSRF( SurfaceType Srf )

Symbolically computes a vector field surface representing the non-normalized normals of the given surface. That is the normal surface, evaluated at (u, v), provides a vector in the direction of the normal of the original surface at (u, v). The normal surface is computed as the symbolic cross product of the two surfaces representing the partial derivatives of the original surface.

Example:

NrmlSrf = SNRMLSRF( Srf );

SPHERE

PolygonType SPHERE( VectorType Center, NumericType Radius )

Creates a SPHERE geometric object, defined by Center as the center of the SPHERE, and with Radius as the radius of the SPHERE. See RESOLUTION for accuracy of SPHERE approximation as a polygonal model. See IRITSTATE's "PrimRatSrfs" and "PrimRatSrfs" state variables.

SRAISE

SurfaceType SRAISE( SurfaceType Srf, ConstantType Direction, NumericType NewOrder )

Raises Srf to the specified NewOrder in the specified Direction.

Example:

Srf = ruledSrf( cbezier( list( ctlpt( E3, -0.5, -0.5, 0.0 ), ctlpt( E3, 0.5, -0.5, 0.0 ) ) ), cbezier( list( ctlpt( E3, -0.5, 0.5, 0.0 ), ctlpt( E3, 0.5, 0.5, 0.0 ) ) ) ); Srf = SRAISE( SRAISE( Srf, ROW, 3 ), COL, 3 );

constructs a bilinear flat ruled surface and raises both its directions to be a bi-quadratic surface.

SREFINE

SurfaceType SREFINE( SurfaceType Srf, ConstantType Direction, NumericType Replace, ListType KnotList )

Provides the ability to Replace a knot vector of Srf or refine it in the specified direction Direction (ROW or COL). KnotList is a list of knots to refine Srf at. All knots should be contained in the parametric domain of Srf in Direction. If the knot vector is replaced, the length of KnotList should be identical to the length of the original knot vector of Srf in Direction. If Srf is a Bezier surface, it is automatically promoted to be a Bspline surface.

Example:

Srf = SREFINE( SREFINE( Srf, ROW, FALSE, list( 0.333, 0.667 ) ), COL, FALSE, list( 0.333, 0.667 ) );

refines Srf in both directions by adding two more knots at 0.333 and 0.667.

SREGION

SurfaceType SREGION( SurfaceType Srf, ConstantType Direction, NumericType MinParam, NumericType MaxParam )

Extracts a region of Srf between MinParam and MaxParam in the specified Direction. Both MinParam and MaxParam should be contained in the parametric domain of Srf in Direction.

Example:

Srf = ruledSrf( cbezier( list( ctlpt( E3, -0.5, -0.5, 0.5 ), ctlpt( E3, 0.0, 0.5, 0.0 ), ctlpt( E3, 0.5, -0.5, 0.0 ) ) ), cbezier( list( ctlpt( E3, -0.5, 0.5, 0.0 ), ctlpt( E3, 0.0, 0.0, 0.0 ), ctlpt( E3, 0.5, 0.5, 0.5 ) ) ) ); SubSrf = SREGION( Srf, ROW, 0.3, 0.6 );

extracts the region of Srf from the parameter value 0.3 to the parameter value 0.6 along the ROW direction. the COLumn direction is extracted as a whole.

SREPARAM

SurfaceType SREPARAM( SurfaceType Srf, ConstantType Direction, NumericType MinParam, NumericType MaxParam )

Reparametrize Srf over a new domain from MinParam to MaxParam, in the prescribed Direction. This operation does not affect the geometry of the surface and only affine transforms its knot vectors. A Bezier surface will automatically be promoted into a Bspline surface by this function.

Example:

srf = sbspline( 2, 4, list( list( ctlpt( E3, 0.0, 0.0, 1.0 ), ctlpt( E2, 0.0, 1.0 ), ctlpt( E3, 0.0, 2.0, 1.0 ) ), list( ctlpt( E2, 1.0, 0.0 ), ctlpt( E3, 1.0, 1.0, 2.0 ), ctlpt( E2, 1.0, 2.0 ) ), list( ctlpt( E3, 2.0, 0.0, 2.0 ), ctlpt( E2, 2.0, 1.0 ), ctlpt( E3, 2.0, 2.0, 2.0 ) ), list( ctlpt( E2, 3.0, 0.0 ), ctlpt( E3, 3.0, 1.0, 2.0 ), ctlpt( E2, 3.0, 2.0 ) ), list( ctlpt( E3, 4.0, 0.0, 1.0 ), ctlpt( E2, 4.0, 1.0 ), ctlpt( E3, 4.0, 2.0, 1.0 ) ) ), list( list( KV_OPEN ), list( KV_OPEN ) ) ); srf = sreparam( sreparam( srf, ROW, 0, 1 ), COL, 0, 1 );

Ensures that the Bspline surface is defined over the unit size parametric domain.

SRINTER

PointType SRINTER( SurfaceType Srf, PointType RayOrigin, VectorType RayDirection )

Computes the first intersection, if any, of the prescribed ray originating from RayOrigin in direction RayDirection with surface Srf. Returns the intersection point in the parametric space of Srf with the U and V coordinates as the X and Y coefficients of the returned value. The intersection is computed between the ray and a polygonal approximation of the surface Srf as set via the RESOLUTION variable.

Example:

RayOrigin = point( 2, 0.1, 0.3 ); RayDir = vector( -4, 0, 0 ); RayLine = coerce( RayOrigin, E3 ) + coerce( RayOrigin + RayDir, E3 ); color( RayLine, magenta ); attrib( RayLine, "dwidth", 2 ); resolution = 5; InterPt = SRINTER( glass, RayOrigin, RayDir ); InterPtE3 = seval( glass, coord( InterPt, 0 ), coord( InterPt, 1 ) ); color( InterPtE3, cyan ); attrib( InterPtE3, "dwidth", 3 ); view( list( InterPtE3, RayLine, glass, axes ), 1 ); resolution = 80; InterPt = SRINTER( glass, RayOrigin, RayDir ); InterPtE3 = seval( glass, coord( InterPt, 0 ), coord( InterPt, 1 ) ); color( InterPtE3, cyan ); attrib( InterPtE3, "dwidth", 3 ); view( list( InterPtE3, RayLine, glass, axes ), 1 );

A complete example of constructing a ray and intersecting it against a surface of a glass at two different resolution, resulting in two different accuracies. See also RESOLUTION.

STANGENT

VectorType STANGENT( SurfaceType Srf, ConstantType Direction, NumericType UParam, NumericType VParam ) or VectorType STANGENT( TrimSrfType Srf, ConstantType Direction, NumericType UParam, NumericType VParam )

Computes the tangent vector to (possibly trimmed) surface Srf at the parameter values UParam and VParam in Direction. The returned vector has a unit length.

Example:

Tang = STANGENT( Srf, ROW, 0.5, 0.6 );

computes the tangent to Srf in the ROW direction at the parameter values (0.5, 0.6).

STRIMSRF

SurfaceType STRIMSRF( TrimSrfType TSrf )

Extracts the surface of a trimmed surface TSrf.

Example:

Srf = STRIMSRF( TrimSrf );

extracts the surface of TrimSrf.

STRIVAR

SurfaceType STRIVAR( TrivarType TV )

Extracts an iso surface from a trivariate function TV.

Example:

TV1 = tbezier( list( list( list( ctlpt( E3, 0.1, 0.0, 0.8 ), ctlpt( E3, 0.2, 0.1, 2.4 ) ), list( ctlpt( E3, 0.3, 2.2, 0.2 ), ctlpt( E3, 0.4, 2.3, 2.0 ) ) ), list( list( ctlpt( E3, 2.4, 0.8, 0.1 ), ctlpt( E3, 2.2, 0.7, 2.3 ) ), list( ctlpt( E3, 2.3, 2.6, 0.5 ), ctlpt( E3, 2.1, 2.5, 2.7) ) ) ) ); Srf = STRIVAR( TV1, col, 0.4 );

extracts an iso surface of TV1, in the col direction at parameter value 0.4.

SURFPREV

SurfaceType SURFPREV( CurveType Object )

Same as SURFREV but approximates the surface of revolution as a polynomial surface. Object must be a polynomial curve. See SURFREV.

SURFREV

PolygonType SURFREV( PolygonType Object ) or SurfaceType SURFREV( CurveType Object )

Creates a surface of revolution by rotating the first polygon/curve of the given Object, around the Z axis. Use the linear transformation function to position a surface of revolution in a different orientation.

Example:

VTailAntn = SURFREV( ctlpt( E3, 0.001, 0.0, 1.0 ) + ctlpt( E3, 0.01, 0.0, 1.0 ) + ctlpt( E3, 0.01, 0.0, 0.8 ) + ctlpt( E3, 0.03, 0.0, 0.7 ) + ctlpt( E3, 0.03, 0.0, 0.3 ) + ctlpt( E3, 0.001, 0.0, 0.0 ) );

constructs a piecewise linear Bspline curve in the XZ plane and uses it to construct a surface of revolution by rotating it around the Z axis. See also SURFPREV.

SVISIBLE

ListType SVISIBLE( SurfaceType Srf, NumericType Resolution, NumericType ConeSize )

Computes a decomposition of a freeform surface Srf into regions each visible with a cone visibility of ConeSize degrees from one direction. In other words all point in one region have angular deviation of their surface normal of less than ConeSize degrees from the set viewing direction. Resolution controls the accuracy of the computation; the higher this value is, more exact the result. 20 is a good starting value. Each returned region is a trimmed surface that has a "ViewDir" attribute that contains the viewing direction of this region.

Example:

c1 = cbezier( list( ctlpt( E3, 1.0, 0.0, 0.5 ), ctlpt( E3, 1.1, 0.0, 0.0 ), ctlpt( E3, 1.0, 0.0, -0.5 ) ) ); Simp = sregion( surfPRev( c1 ), col, 0.0, 1.0 ) * rz( 45 ) * rx( 90 ); Decomp = SVISIBLE( Simp, 20, 30 * pi / 180 ); SimDecomp = nil(); Mod = 5; for ( i = 1, 1, sizeof( Decomp ), o = nth( Decomp, i ): v = getattr( o, "ViewDir" ): l = ( ctlpt( E3, 0, 0, 0 ) + coerce( v, e3 ) ) * sc( 1.5 ): j = floor( ( i - 1 ) / Mod ): snoc( list( o, Simp, l, axes ) * view_mat * tx( ( i - 1 - j * Mod ) * 2 - 4 ) * ty( -j * 2 ), SimDecomp ) ); view( SimDecomp, on );

decomposes a given surface Simp into regions of 30 degrees at most, goes over the decomposed regions and order them five in a row.

SWEEPSRF

SurfaceType SWEEPSRF( CurveType CrossSection, CurveType Axis, CurveType FrameCrv | VectorType FrameVec | ConstType OFF )

Constructs a generalized cylinder surface. This function sweeps a specified cross-section CrossSection along the provided Axis. By default, when frame specification is OFF, the orientation of the cross section is computed using the Axis curve tangent and normal. However, unlike the Frenet frame, attempt is made to minimize the normal change, as can happen along inflection points in Axis. If a VectorType FrameVec is provided as a frame orientation setting, it is used to fix the binormal direction to this value. In other words, the orientation frame has a fixed binormal. If a CurveType FrameCrv is specified as a frame orientation setting, this vector field curve is evaluated at each placement of the cross-section to yield the needed binormal.

The resulting sweep is only an approximation of the real sweep. The resulting sweep surface will not be exact, in general. Refinement of the axis curve at the proper location, where accuracy is important, should improve the accuracy of the output. The parametric domains of FrameCrv do not have to match the parametric domain of Axis, and its parametric domain is automatically made compatible by this function.

Example:

Cross = arc( vector( 0.2, 0.0, 0.0 ), vector( 0.2, 0.2, 0.0 ), vector( 0.0, 0.2, 0.0 ) ) + arc( vector( 0.0, 0.4, 0.0 ), vector( 0.1, 0.4, 0.0 ), vector( 0.1, 0.5, 0.0 ) ) + arc( vector( 0.8, 0.5, 0.0 ), vector( 0.8, 0.3, 0.0 ), vector( 1.0, 0.3, 0.0 ) ) + arc( vector( 1.0, 0.1, 0.0 ), vector( 0.9, 0.1, 0.0 ), vector( 0.9, 0.0, 0.0 ) ) + ctlpt( E2, 0.2, 0.0 ); Axis = arc( vector( -1.0, 0.0, 0.0 ), vector( 0.0, 0.0, 0.1 ), vector( 1.0, 0.0, 0.0 ) ); Axis = crefine( Axis, FALSE, list( 0.25, 0.5, 0.75 ) ); Srf1 = SWEEPSRF( Cross, Axis, OFF ); Srf2 = SWEEPSRF( Cross, Axis, vector( 0.0, 1.0, 1.0 ) ); Srf3 = SWEEPSRF( Cross, Axis, cbezier( list( ctlpt( E3, 1.0, 0.0, 0.0 ), ctlpt( E3, 0.0, 1.0, 0.0 ), ctlpt( E3, -1.0, 0.0, 0.0 ) ) ) );

constructs a rounded rectangle cross-section and sweeps it along an arc, while orienting it several ways. The axis curve Axis is manually refined to better approximate the requested shape.

See also SWPSCLSRF for sweep with scale.

SWPSCLSRF

SurfaceType SWPSCLSRF( CurveType CrossSection, CurveType Axis, NumericType Scale | CurveType ScaleCrv, CurveType FrameCrv | VectorType FrameVec | ConstType OFF, NumericType ScaleRefine )

Constructs a generalized cylinder surface. This function sweeps a specified cross-section CrossSection along the provided Axis. The cross-section may be scaled by a constant value Scale, or scaled along the Axis parametric direction via a scaling curve ScaleCrv. By default, when frame specification is OFF, the orientation of the cross section is computed using the Axis curve tangent and normal. However, unlike the Frenet frame, attempt is made to minimize the normal change, as can happen along inflection points in Axis. If a VectorType FrameVec is provided as a frame orientation setting, it is used to fix the binormal direction to this value. In other words, the orientation frame has a fixed binormal. If a CurveType FrameCrv is specified as a frame orientation setting, this vector field curve is evaluated at each placement of the cross-section to yield the needed binormal. ScaleRefine is an integer value to define possible refinement of the Axis to reflect the information in ScalingCrv. Value of zero will force no refinement while value of n > 0 will insert n times the number of control points in ScaleCrv into Axis, better emulating the scaling requested. The resulting sweep is only an approximation of the real sweep. The scaling and axis placement will not be exact, in general. Manual refinement (in addition to ScaleRefine) of the axis curve at the proper location, where accuracy is important, should improve the accuracy of the output. The parametric domains of ScaleCrv and FrameCrv do not have to match the parametric domain of Axis, and their domains are made compatible by this function.

Example:

Cross = arc( vector( -0.11, -0.1, 0.0 ), vector( -0.1, -0.1, 0.0 ), vector( -0.1, -0.11, 0.0 ) ) + arc( vector( 0.1, -0.11, 0.0 ), vector( 0.1, -0.1, 0.0 ), vector( 0.11, -0.1, 0.0 ) ) + arc( vector( 0.11, 0.1, 0.0 ), vector( 0.1, 0.1, 0.0 ), vector( 0.1, 0.11, 0.0 ) ) + arc( vector( -0.1, 0.11, 0.0 ), vector( -0.1, 0.1, 0.0 ), vector( -0.11, 0.1, 0.0 ) ) + ctlpt( E2, -0.11, -0.1 ); scaleCrv = cbspline( 3, list( ctlpt( E2, 0.05, 1.0 ), ctlpt( E2, 0.1, 0.0 ), ctlpt( E2, 0.2, 2.0 ), ctlpt( E2, 0.3, 0.0 ), ctlpt( E2, 0.4, 2.0 ), ctlpt( E2, 0.5, 0.0 ), ctlpt( E2, 0.6, 2.0 ), ctlpt( E2, 0.7, 0.0 ), ctlpt( E2, 0.8, 2.0 ), ctlpt( E2, 0.85, 1.0 ) ), list( KV_OPEN ) ); Axis = circle( vector( 0, 0, 0 ), 1 ); Frame = circle( vector( 0, 0, 0 ), 1 ) * rotx( 90 ) * trans( vector( 1.5, 0.0, 0.0 ) ); Srf1 = SWPSCLSRF( Cross, Axis, scaleCrv, off, 0 ); Srf2 = SWPSCLSRF( Cross, Axis, scaleCrv, off, 2 ); Srf3 = SWPSCLSRF( Cross, Axis, 1.0, Frame, 0 );

constructs a rounded rectangle cross-section and sweeps it along a circle, while scaling and orienting in several ways. The axis curve Axis is automatically refined in Srf2 to better approximate the requested scaling.

See also SWEEPSRF for sweep with no scale.

SYMBPROD

CurveType SYMBPROD( CurveType Crv1, CurveType Crv2 ) or SurfaceType SYMBPROD( SurfaceType Srf1, SurfaceType Srf2 )

Computes the symbolic product of the given two curves or surfaces as a curve or surface. The product is computed coordinate-wise.

Example:

ProdSrf = SYMBPROD( Srf1, Srf2 )

SYMBDPROD

CurveType SYMBDPROD( CurveType Crv1, CurveType Crv2 ) or SurfaceType SYMBDPROD( SurfaceType Srf1, SurfaceType Srf2 )

Computes the symbolic dot (inner) product of the given two curves or surfaces as a scalar curve or surface.

Example:

DiffCrv = symbdiff( Crv1, Crv2 ) DistSqrCrv = SYMBDPROD( DiffCrv, DiffCrv )

Computes a scalar curve that at parameter t is equal to the distance square between Crv1 at t and Crv2.

SYMBCPROD

CurveType SYMBCPROD( CurveType Crv1, CurveType Crv2 ) or SurfaceType SYMBCPROD( SurfaceType Srf1, SurfaceType Srf2 )

Computes the symbolic cross product of the given two curves or surfaces as a curve or surface.

Example:

NrmlSrf = SYMBCPROD( sderive( Srf, ROW ), sderive( Srf, COL ) )

computes a normal surface as the cross product of the surface two partial derivatives (see SNRMLSRF).

SYMBSUM

CurveType SYMBSUM( CurveType Crv1, CurveType Crv2 ) or SurfaceType SYMBSUM( SurfaceType Srf1, SurfaceType Srf2 )

Computes the symbolic sum of the given two curves or surfaces as a curve or surface. The sum is computed coordinate-wise.

Example:

SumCrv = SYMBSUM( Crv1, Crv2 )

SYMBDIFF

CurveType SYMBDIFF( CurveType Crv1, CurveType Crv2 ) or SurfaceType SYMBDIFF( SurfaceType Srf1, SurfaceType Srf2 )

Computes the symbolic difference of the given two curves or surfaces as a curve or surface. The difference is computed coordinate-wise.

Example:

DiffCrv = SYMBDIFF( Crv1, Crv2 ) DistSqrCrv = symbdprod( DiffCrv, DiffCrv )

TBEZIER

TrivarType TBEZIER( ListType CtlMesh )

Creates a Bezier trivariate using the provided control mesh. CtlMesh is a list of planes, each of which is a list of rows, each of which is a list of control points. All control points must be of type (E1-E5, P1-P5), or regular PointType defining the trivariate's control mesh. Surface's point type will be of a space which is the union of the spaces of all points.

Example:

TV = TBEZIER( list( list( list( ctlpt( E3, 0.1, 0.1, 0.0 ), ctlpt( E3, 0.2, 0.5, 1.1 ), ctlpt( E3, 0.3, 0.1, 2.2 ) ), list( ctlpt( E3, 0.4, 1.3, 0.5 ), ctlpt( E3, 0.5, 1.7, 1.7 ), ctlpt( E3, 0.6, 1.3, 2.9 ) ), list( ctlpt( E3, 0.7, 2.4, 0.5 ), ctlpt( E3, 0.8, 2.6, 1.4 ), ctlpt( E3, 0.9, 2.8, 2.3 ) ) ), list( list( ctlpt( E3, 1.1, 0.1, 0.5 ), ctlpt( E3, 1.3, 0.2, 1.7 ), ctlpt( E3, 1.5, 0.3, 2.9 ) ), list( ctlpt( E3, 1.7, 1.2, 0.0 ), ctlpt( E3, 1.9, 1.4, 1.2 ), ctlpt( E3, 1.2, 1.6, 2.4 ) ), list( ctlpt( E3, 1.4, 2.3, 0.9 ), ctlpt( E3, 1.6, 2.5, 1.7 ), ctlpt( E3, 1.8, 2.7, 2.5 ) ) ) ) );

creats a trivariate Bezier which is linear in the first direction, and quadratic in the second and third.

TBSPLINE

TrivarType TBSPLINE( NumericType UOrder, NumericType VOrder, NumericType WOrder, ListType CtlMesh, ListType KnotVectors )

Creates a Bspline trivariate with the provided UOrder, VOrder and WOrder orders, the control mesh CtlMesh, and the three knot vectors in KnotVectors. CtlMesh is a list of planes, each of which is a list of rows, each of which is a list of control points. All control points must be of point type (E1-E5, P1-P5), or regular PointType defining the trivariate's control mesh. Trivariate's point type will be of a space which is the union of the spaces of all points. KnotVectors is a list of three knot vectors. Each knot vector is a list of NumericType knots of length #CtlPtList plus the Order. If, however, the length of the knot vector is equal to #CtlPtList + Order + Order - 1} the curve is assumed periodic. The knot vector may also be a list of a single constant KV_OPEN or KV_FLOAT or KV_PERIODIC, in which a uniform knot vector with the appropriate length and with open, floating or periodic end condition will be constructed automatically.

Example:

TV = TBSPLINE( 2, 2, 2, list( list( list( ctlpt( E3, 0.1, 0.1, 0.0 ), ctlpt( E3, 0.2, 0.5, 1.1 ), ctlpt( E3, 0.3, 0.1, 2.2 ) ), list( ctlpt( E3, 0.4, 1.3, 0.5 ), ctlpt( E3, 0.5, 1.7, 1.7 ), ctlpt( E3, 0.6, 1.3, 2.9 ) ), list( ctlpt( E3, 0.7, 2.4, 0.5 ), ctlpt( E3, 0.8, 2.6, 1.4 ), ctlpt( E3, 0.9, 2.8, 2.3 ) ) ), list( list( ctlpt( E3, 1.1, 0.1, 0.5 ), ctlpt( E3, 1.3, 0.2, 1.7 ), ctlpt( E3, 1.5, 0.3, 2.9 ) ), list( ctlpt( E3, 1.7, 1.2, 0.0 ), ctlpt( E3, 1.9, 1.4, 1.2 ), ctlpt( E3, 1.2, 1.6, 2.4 ) ), list( ctlpt( E3, 1.4, 2.3, 0.9 ), ctlpt( E3, 1.6, 2.5, 1.7 ), ctlpt( E3, 1.8, 2.7, 2.5 ) ) ) ), list( list( KV_OPEN ), list( KV_OPEN ), list( KV_OPEN ) ) );

constructs a trilinear Bspline trivariate with open end conditions. TCRVTR

TCRVTR

AnyType TCRVTR( TrivarType TV, PointType Pos, NumericType ComputeWhat )

Compute differential curvature properties of an isosurface of the given trivariate TV at the given (parameteric) location Pos. Following the value of ComputeWhat, the result equals,

-1	Initialization (a must prelude)
0	Conclusion (a must postlude)
1	Returns a vector hold the gradient
2	Returns a list of three vectors
	equal to the Hessian of this location
3	Returns a list of two scalar values
	(Principal curvatures) and two vectors
	(Principal directions).

Every evaluation must start with an invokation with ComputeWhat equal -1 and terminates with ComputeWhat 0. In both cases, 1 is return in case of success.

Example:

TCRVTR( TV, point( 0, 0, 0 ), -1 ); # Prelude Grad1 = TCRVTR( TV, point( 0, 0, 0 ), 1 ); Grad2 = TCRVTR( TV, point( 0, 0, 1 ), 1 ); Grad3 = TCRVTR( TV, point( 0, 1, 0 ), 1 ); Grad4 = TCRVTR( TV, point( 1, 0, 0 ), 1 ); TCRVTR( TV, point( 0, 0, 0 ), 0 ); #Postlude

TDERIVE

TrivarType TDERIVE( TrivarType TV, NumericType Dir )

Returns a vector field trivariate representing the differentiated trivariate in the given direction (ROW, COL, or DEPTH). Evaluation of the returned trivariate at a given parameter value will return a vector representing the partial derivative of TV in Dir at that parameter value.

TV = tbezier( list( list( list( ctlpt( E1, 0.1 ), ctlpt( E1, 0.2 ) ), list( ctlpt( E1, 0.3 ), ctlpt( E1, 0.4 ) ) ), list( list( ctlpt( E1, 2.4 ), ctlpt( E1, 2.2 ) ), list( ctlpt( E1, 2.3 ), ctlpt( E1, 2.1 ) ) ) ) ); DuTV = TDERIVE( TV, ROW ); DvTV = TDERIVE( TV, COL ); DwTV = TDERIVE( TV, DEPTH );

computes the gradiate of a scalar trivariate field, by computing its partials with respect to u, v, and w.

TDIVIDE

TrivarType TDIVIDE( TrivarType TV, ConstantType Direction, NumericType Param )

Subdivides a trivariate into two at the specified parameter value Param in the specified Direction (ROW, COL, or DEPTH). TV can be either a Bspline trivairate in which Param must be conatined in the parametric domain of the trivariate, or a Bezier trivariate in which Param must be in the range of zero to one.

It returns a list of the two sub-trivariates. The individual trivariates may be extracted from the list using the NTH command.

Example:

TvDiv = TDIVIDE( Tv2, depth, 0.3 ); Tv2a = nth( TvDiv, 1 ) * tx( -2.2 ); Tv2b = nth( TvDiv, 2 ) * tx( 2.0 );

subdivides Tv2 at the parameter value of 0.3 in the DEPTH direction,

TEDITPT

TrivarType TEDITPT( TrivarType TV, CtlPtType CPt, NumericType UIndex, NumericType VIndex ) NumericType WIndex )

Provides a simple mechanism to manually modify a single control point number UIndex, VIndex and WIndex (base count is 0) in the control mesh of Srf by substituting CtlPt instead. CtlPt must have the same point type as the control points of Srf. Original surface Srf is not modified.

Example:

CPt = ctlpt( E3, 1, 2, 3 ); NewTV = TEDITPT( TV, CPt, 0, 0, 0 );

Constructs a NewTV with the first control point of TV being CPt.

TEVAL

CtlPtType TEVAL( TrivarType TV, NumericType UParam, NumericType VParam, NumericType WParam )

Evaluates the provided trivariate TV at the given UParam, VParam and WParam values. UParam, VParam, WParam must be contained in the surface parametric domain if TV is a Bspline surface, or between zero and one if TV is a Bezier trivariate. The returned control point has the same type as the control points of TV.

Example:

CPt = TEVAL( TV1, 0.25, 0.22, 0.7 );

evaluates TV at the parameter values of (0.25, 0.22, 0.7).

TEXTGEOM

AnyType TEXTGEOM( StringType Str, VectorType Spacing, NumericType Scaling )

Creates a displayable geometry that represents the text in Str, with Spacing space between individual characters. Each character is scaled by Scaling where scaling of one generates a close to a unit size character.

Example:

a = TEXTGEOM("Text", vector( 0.12, 0, 0 ), 0.1 ); b = TEXTGEOM("IRIT", vector( 0, -0.12, 0 ), 0.1 );

Creates an horizontal Text and a vertical top to bottom IRIT, both as geometrical objects.

TFROMSRFS

TrivarType TFROMSRFS( ListType SrfList, NumericType OtherOrder )

Constructs a trivariate by substituting the surfaces in SrfList as planes in a control mesh of a trivariate. Surfaces in SrfList are made compatible by promoting Bezier surfaces to Bsplines if necessary, and raising degree and refining as required before substituting the control meshes of the surfaces as planes in the mesh of the trivariate. The other, third, direction order is set by OtherOrder, which cannot be larger than the number of surfaces.

The trivariate interpolates the first and last surfaces only.

Example:

s1 = sbezier( list( list( ctlpt( E3, -0.5, -0.5, 0 ), ctlpt( E3, -0.5, 0.5, 0 ) ), list( ctlpt( E3, 0.5, -0.5, 0 ), ctlpt( E3, 0.5, 0.5, 0 ) ) ) ) * sc( 0.3 ); Srfs = list( s1 * sc( 2.0 ), s1 * sx( 1.4 ) * ry( 45 ) * tz( 1.0 ), s1 * ry( 90 ) * trans( vector( 1.0, 0.0, 1.1 ) ), s1 * sx( 1.4 ) * ry( 135 ) * trans( vector( 2.0, 0.0, 1.0 ) ), s1 * sc( 2.0 ) * ry( 180 ) * trans( vector( 2.0, 0.0, 0.0 ) ) ); color( Srfs, red ); ts = tfromsrfs( Srfs, 3 ); color( ts, green ); view( list( Srfs, ts ), on );

Constructs a trivariate from five planar surfaces and display both the trivariate and the five planar surfaces, in different colors.

TINTERP

TrivarType TINTERP( TrivarType TV );

Given a trivariate data structure, computes a new trivariate in the same function space (i.e. same knot sequences and orders) that interpolates the given triavriate, TV, at the node parameter values.

Example:

tv = tbspline( 3, 3, 2, list( list( list( ctlpt( E3, 0.1, 0.1, 0.0 ), ctlpt( E3, 0.2, 0.5, 1.1 ), ctlpt( E3, 0.3, 0.1, 2.2 ) ), list( ctlpt( E3, 0.4, 1.3, 0.5 ), ctlpt( E3, 0.5, 1.7, 1.7 ), ctlpt( E3, 0.6, 1.3, 2.9 ) ), list( ctlpt( E3, 0.7, 2.4, 0.5 ), ctlpt( E3, 0.8, 2.6, 1.4 ), ctlpt( E3, 0.9, 2.8, 2.3 ) ) ), list( list( ctlpt( E3, 1.1, 0.1, 0.5 ), ctlpt( E3, 1.3, 0.2, 1.7 ), ctlpt( E3, 1.5, 0.3, 2.9 ) ), list( ctlpt( E3, 1.7, 1.2, 0.0 ), ctlpt( E3, 1.9, 1.4, 1.2 ), ctlpt( E3, 1.2, 1.6, 2.4 ) ), list( ctlpt( E3, 1.4, 2.3, 0.9 ), ctlpt( E3, 1.6, 2.5, 1.7 ), ctlpt( E3, 1.8, 2.7, 2.5 ) ) ), list( list( ctlpt( E3, 2.8, 0.1, 0.4 ), ctlpt( E3, 2.6, 0.7, 1.3 ), ctlpt( E3, 2.4, 0.2, 2.2 ) ), list( ctlpt( E3, 2.2, 1.1, 0.4 ), ctlpt( E3, 2.9, 1.2, 1.5 ), ctlpt( E3, 2.7, 1.3, 2.6 ) ), list( ctlpt( E3, 2.5, 2.9, 0.7 ), ctlpt( E3, 2.3, 2.8, 1.7 ), ctlpt( E3, 2.1, 2.7, 2.7 ) ) ) ), list( list( KV_OPEN ), list( KV_OPEN ), list( KV_OPEN ) ) ); tvi = TINTERP( tv );

creates a quadratic by quaratic by linear trivairatiate tvi that interpolates the control points of tv at the node parameter values.

TMORPH

TrivarType TMORPH( TrivarType TV1, SurfaceType TV2, NumericType Blend )

Creates a new trivariate which is a convex blend of the two given trivariates. The two given trivariates must be compatible (see FFCOMPAT) before this blend is invoked. Very useful if a sequence that "morphs" one trivariate to another is to be created and in combination of MRCHCUBE.

Example:

Size = 0.05; for ( i = 0, step, 1.0, Tv = TMORPH( Tv1, Tv2, i ): view( mrchcube( list( Tv, 1, off ), point( Size, Size, Size ), 1, IsoVal ), on ) );

creates a sequence of 1/step trivariates, morphed from Tv1 to Tv2 and displays an extracted iso surface at level IsoVal. See also MRCHCUBE, CMORPH and SMORPH.

TORUS

PolygonType TORUS( VectorType Center, VectorType Normal, NumericType MRadius, NumericType mRadius )

Creates a TORUS geometric object, defined by Center as the center of the TORUS, Normal as the normal to the main plane of the TORUS, MRadius and mRadius as the major and minor radii of the TORUS. See RESOLUTION for the accuracy of the TORUS approximation as a polygonal model. See IRITSTATE's "PrimRatSrfs" and "PrimRatSrfs" state variables.

Example:

T = TORUS( vector( 0.0, 0.0, 0.0), vector( 0.0, 0.0, 1.0), 0.5, 0.2 );

constructs a torus with major plane as the XY plane, major radius of 0.5, and minor radius of 0.2.

TRAISE

TrivarType TRAISE( TrivarType TV, ConstantType Direction, NumericType NewOrder )

Raise TV to the specified NewOrder in the specified Direction.

Example:

tv1r = TRAISE( traise( traise( tv1, row, 4 ), col, 4 ), depth, 4 );

Ensures that the trivariate is a tricubic.

TREFINE

TrivarType TREFINE( TrivarType TV, ConstantType Direction, NumericType Replace, ListType KnotList )

Provides the ability to Replace a knot vector of TV or refine it in the specified direction Direction (ROW, COL, or DEPTH). KnotList is a list of knots to refine TV at. All knots should be contained in the parametric domain of TV in Direction. If the knot vector is replaced, the length of KnotList should be identical to the length of the original knot vector of TV in Direction. If TV is a Bezier trivariate, it is automatically promoted to be a Bspline trivariate.

Example:

TV = TREFINE( TREFINE( TREFINE( TV, ROW, FALSE, list( 0.333, 0.667 ) ), COL, FALSE, list( 0.333, 0.667 ) ), DEPTH, FALSE, list( 0.333, 0.667 ) );

refines TV in all directions by adding two more knots at 0.333 and 0.667.

TREGION

TrivarType TREGION( TrivarType TV, ConstantType Direction, NumericType MinParam, NumericType MaxParam )

Extracts a region of TV between MinParam and MaxParam in the specified Direction. Both MinParam and MaxParam should be contained in the parametric domain of TV in Direction.

Example:

Tv1 = tbezier( list( list( list( ctlpt( E3, 0.1, 0.0, 0.8 ), ctlpt( E3, 0.2, 0.1, 2.4 ) ), list( ctlpt( E3, 0.3, 2.2, 0.2 ), ctlpt( E3, 0.4, 2.3, 2.0 ) ) ), list( list( ctlpt( E3, 2.4, 0.8, 0.1 ), ctlpt( E3, 2.2, 0.7, 2.3 ) ), list( ctlpt( E3, 2.3, 2.6, 0.5 ), ctlpt( E3, 2.1, 2.5, 2.7) ) ) ) ); Tv1r1 = TREGION( Tv1, row, 0.1, 0.2 ); Tv1r2 = TREGION( Tv1, row, 0.4, 0.6 ); Tv1r3 = TREGION( Tv1, row, 0.99, 1.0 );

extracts three regions of Tv1 along the ROW direction.

TREPARAM

TrivarType TREPARAM( TrivarType TV, ConstantType Direction, NumericType MinParam, NumericType MaxParam )

Reparametrize TV over a new domain from MinParam to MaxParam, in the prescribed Direction. This operation does not affect the geometry of the trivariate and only affine transforms its knot vectors. A Bezier trivariate will automatically be promoted into a Bspline surface by this function.

Example:

Tv = TREPARAM( TREPARAM( TREPARAM( tv, row, 0, 1 ), col, 0, 1 ), depth, 0, 1 );

Ensures that the trivariate is defined over the unit size parametric cube.

TRIANGL

PolygonType TRIANGL( PolygonType Model )

Converts Model into a new model with exactly the same shape that holds only triangles.

TRIMSRF

TrimSrfType TRIMSRF( SurfaceType Srf, CurveType TrimCrv, NumericType HasUpperLevel ) or TrimSrfType TRIMSRF( SurfaceType Srf, ListType TrimCrvs, NumericType HasUpperLevel )

Creates a trimmed surface from the provided surface Srf and the trimming curve TrimCrv or curves TrimCrvs. If HasUpperLevel is FALSE, an additional trimming curve is automatically being added that contains the entire parametric domain of Srf. No validity test is performed on the trimming curves which are assumed two dimensional curves contained in the parametric domain of Srf.

Example:

spts = list( list( ctlpt( E3, 0.1, 0.0, 1.0 ), ctlpt( E3, 0.3, 1.0, 0.0 ), ctlpt( E3, 0.0, 2.0, 1.0 ) ), list( ctlpt( E3, 1.1, 0.0, 0.0 ), ctlpt( E3, 1.3, 1.5, 2.0 ), ctlpt( E3, 1.0, 2.1, 0.0 ) ), list( ctlpt( E3, 2.1, 0.0, 2.0 ), ctlpt( E3, 2.3, 1.0, 0.0 ), ctlpt( E3, 2.0, 2.0, 2.0 ) ), list( ctlpt( E3, 3.1, 0.0, 0.0 ), ctlpt( E3, 3.3, 1.5, 2.0 ), ctlpt( E3, 3.0, 2.1, 0.0 ) ), list( ctlpt( E3, 4.1, 0.0, 1.0 ), ctlpt( E3, 4.3, 1.0, 0.0 ), ctlpt( E3, 4.0, 2.0, 1.0 ) ) ); sb = sbspline( 3, 3, spts, list( list( KV_OPEN ), list( KV_OPEN ) ) ); TCrv1 = cbspline( 2, list( ctlpt( E2, 0.3, 0.3 ), ctlpt( E2, 0.7, 0.3 ), ctlpt( E2, 0.7, 0.7 ), ctlpt( E2, 0.3, 0.7 ), ctlpt( E2, 0.3, 0.3 ) ), list( KV_OPEN ) ); TCrv2 = circle( vector( 0.5, 0.5, 0.0 ), 0.25 ); TCrv3 = cbspline( 3, list( ctlpt( E2, 0.3, 0.3 ), ctlpt( E2, 0.7, 0.3 ), ctlpt( E2, 0.7, 0.7 ), ctlpt( E2, 0.3, 0.7 ) ), list( KV_PERIODIC ) ); TSrf1 = TRIMSRF( sb, TCrv1, false ); TSrf2 = TRIMSRF( sb, TCrv1, true ); TSrf3 = TRIMSRF( sb, list( TCrv1, TcRv2 * ty( 1 ), TCrv3 * ty( 2 ) ), false );

constructs three trimmed surfaces. Tsrf1 contains the outer boundary and excludes what is inside TCrv1, TSrf2 contains only the domain inside TCrv1. TCrv3 has three holes corresponds to the three trimming curves.

TSBEZIER

SurfaceType TSBEZIER( NumericType Order, ListType CtlMesh )

Creates a triangular Bezier surface of oder Order using the provided control mesh. CtlMesh is a list of control points of size (Order + 1) * Order / 2. All control points must be of type (E1-E5, P1-P5), or regular PointType defining the surface's control polygon. Surface's point type will be of a space which is the union of the spaces of all points.

Example:

b = TSBEZIER( 3, list( ctlpt( E3, 0.0, 0.0, 0.4 ), ctlpt( E3, 0.3, 0.0, 0.3 ), ctlpt( E3, 0.7, 0.0, 0.8 ), ctlpt( E3, 0.2, 0.4, 1.0 ), ctlpt( E3, 0.4, 0.5, 1.0 ), ctlpt( E3, 0.5, 1.0, 0.7 ) ) );

TSBSPLINE

TriSrfType TSBSPLINE( NumericType Order, NumericType Length, ListType CtlMesh, ListType KnotVector )

Creates a Bspline surface from the provided Order and Length, the control mesh CtlMesh, and the knot vector KnotVector. CtlMesh is a list of control points of size (Length + 1) * Length / 2. All control points must be of point type (E1-E5, P1-P5), or regular PointType defining the surface's control mesh. Surface's point type will be of a space which is the union of the spaces of all points. KnotVector is a list of NumericType knots of length Length plus the Order. The knot vector may also be a list of a single constant KV_OPEN or KV_FLOAT, in which a uniform knot vector with the appropriate length and with open or floating end condition will be constructed automatically.

Not fully supported at this time.

TSDERIVE

TriSrfType TSDERIVE( TriSrfType Srf, NumericType Dir )

Returns a vector field surface representing the differentiated triangular surface in the given direction (ROW, COL, or DEPTH). Evaluation of the returned surface at a given parameter value will return a vector tangent to Srf in Dir at that parameter value.

DuSrf = TSDERIVE( Srf, ROW ); DvSrf = TSDERIVE( Srf, COL ); Normal = coerce( tseval( DuSrf, 0.5, 0.25, 0.25 ), VECTOR_TYPE ) ^ coerce( tseval( DvSrf, 0.5, 0.25, 0.25 ), VECTOR_TYPE );

computes two partial derivatives of the surface Srf and computes its normal as their cross product, at the parametric location (0.5, 0.25, 0.25).

See also TSNORMAL

TSEVAL

CtlPtType TSEVAL( TriSrfType Srf, NumericType UParam, NumericType VParam, NumericType WParam )

Evaluates the provided triangulae surface Srf at the given UParam, VParam, WParam parameters. UParam, VParam, and WParam must all be non negative and must sum to one for a Bezier triangular surface or to the maximum domain, if Bspline surface.

Example:

CPt = TSEVAL( Srf, u, v, 1.0 - u - v );

Evaluates Srf at the parameter values prescribed by u and v.

TSNORMAL

VectorType TSNORMAL( TriSrfType Srf, NumericType UParam, NumericType VParam, NumericType WParam)

Computes the normal vector to a triangular surface Srf at the parameter values UParam, VParam, WParam. The returned vector has a unit length.

UParam, VParam, and WParam must all be non negative and must sum to one for a Bezier triangular surface or to the maximum domain, if Bspline surface.

Example:

Normal = TSNORMAL( Srf, 0.5, 0.5, 0.0 );

computes the normal to Srf at the parameter values (0.5, 0.5, 0.0).

TVLOAD

TrivarType TVLOAD( StringType FileName, NumericType DataType, VectorType VolSize, VectorType Orders )

Loads a volumetric data set from file FileName in as a trivariate of orders Orders. DataType can be one of:

1	Regular ascii (separated by while spaces).
2	Two bytes short integer.
3	Four bytes long integer.
4	One byte (char) integer.
5	Four bytes float.
6	Eight bytes double.

Beware of the little vs big Endian problem! We assume here you read in the volume in the same machine type this file was writeen with.

VolSize provides the dimensions of the volume, with width first and depth last. Uniform open end condition knot vector are constructed to all three axes.

Example:

Tv = tvload( "3dhead", 1, vector( 32, 32, 13 ), vector( 3, 3, 3 ) );

loads the data set "3dhead" of size (32, 32, 13) as a triquadratic function. Data set is assumed to contain ascii numeric values.

See also MRCHCUBE.

Object transformation functions

All the routines in this section construct a 4 by 4 homogeneous transformation matrix representing the required transform. These matrices may be concatenated to achieve more complex transforms using the matrix multiplication operator *. For example, the expression

m = trans( vector( -1, 0, 0 ) ) * rotx( 45 ) * trans( vector( 1, 0, 0 ) );

constructs a transform to rotate an object around the X = 1 line, 45 degrees. A matrix representing the inverse transformation can be computed as:

InvM = m ^ -1

See also overloading of the - operator.

HOMOMAT

MatrixType HOMOMAT( ListType MatData )

Creates an arbitrary homogeneous transformation matrix by manually providing its 16 coefficients. Example:

for ( a = 1, 1, 720 / step, view_mat = save_mat * HOMOMAT( list( list( 1, 0, 0, 0 ), list( 0, 1, 0, 0 ), list( 0, 0, 1, -a * step / 500 ), list( 0, 0, 0, 1 ) ) ): view( list( view_mat, axes ), on ) );

looping and viewing through a sequence of perspective transforms, created using the HOMOMAT constructor.

ROTVEC

MatrixType ROTVEC( VectorType Vec, NumericType Angle )

Creates a rotation around the vector Vec matrix with Angle degrees.

ROTX

MatrixType ROTX( NumericType Angle )

Creates a rotation around the X transformation matrix with Angle degrees.

ROTY

MatrixType ROTY( NumericType Angle )

Creates a rotation around te Y transformation matrix with Angle degrees.

ROTZ

MatrixType ROTZ( NumericType Angle )

Creates a rotation around the Z transformation matrix with Angle degrees.

ROTZ2V

MatrixType ROTZ2V( VectorType Dir )

Creates a rotation matrix that takes Z axis into Dir. Length of Dir is ignored.

ROTZ2V2

MatrixType ROTZ2V2( VectorType Dir, VectorType Dir2 )

Creates a rotation matrix that takes Z axis into Dir, while the X axis is aligned with Dir2. The lengths of Dir and Dir2 are ignored.

SCALE

MatrixType SCALE( VectorType ScaleFactors )

Creates a scaling by the ScaleFactors transformation matrix.

TRANS

MatrixType TRANS( VectorType TransFactors )

Creates a translation by the TransFactors transformation matrix.

General purpose functions

ATTRIB

ATTRIB( AnyType Object, StringType Name, AnyType Value )

Provides a mechanism to add an attribute of any type to an Object, with name Name and value Value. This ATTRIB function is tuned and optimized toward numeric values or strings as Value although any other object type can be saved as attribue.

These attributes may be used to pass information to other programs about this object, and are saved with the objects in data files. Attributes placed on a list object or even a whole hierarchy of objects will be propagated into all items in the list or hierarchy. The exception to this propagation are few. The "animation" attribute that is not propagated and is kept in the internal nodes, forming an hierachy of animation commands for all the objects contained in the list/hierarchy. The "invisible" attribute is saved at all levels of the hierarchy, used to denote a complete sub tree that is invisible (yet can serve as a source instances can point at).

For example,

ATTRIB(Glass, "rgb", "255,0,0"); ATTRIB(Glass, "refract", "1.4"); . . . RMATTR(Glass, "rgb"); # Removes "rgb" attribute.

sets the RGB color and refraction index of the Glass object and later removes the RGB attribute.

Attribute names are case insensitive. Spaces are allowed in the Value string, as well as the double quote itself, although the latter must be escaped:

ATTRIB(Glass, "text", "Say "this is me"");

See also RMATTR for removal of attributes as well as AWIDTH, ADWIDTH, and COLOR.

ADWIDTH

ADWIDTH( GeometricType Object, NumericType DWidth )

Sets the width of the object. This display width is used in pixels in display devices for width of line drawing, if supported by the display device. See also ATTRIB, COLOR, and AWIDTH.

This function is equivament to using,

ATTRIB( Object, "dwidth", Width );

AWIDTH

AWIDTH( GeometricType Object, NumericType Width )

Sets the width of the object to one of those specified below. This width is used in real object side dimensions in tools such as scan converters and rendering tools for rendering lines and curves, as well as postscript. See also ATTRIB, COLOR, and ADWIDTH.

This function is equivament to using,

ATTRIB( Object, "width", Width );

CHDIR

CHDIR( StringType NewDir )

Sets the current working directory to be NewDir.

CLNTCLOSE

CLNTCLOSE( NumericType Handler, NumericType Kill )

Closes a communication channel to a client. Handler contains the index of the communication channel opened via CLNTEXEC. If Kill, the client is send an exit request for it to die. Otherwise, the communication is closed and the client is running stand alone. See also CLNTREAD, CLNTWRITE, and CLNTEXEC.

Example:

h2 = clntexec( "nuldrvs -s-" ); . . . CLNTCLOSE( h2,TRUE );

closes the connection to the nuldrvs client, opened via CLNTEXEC.

CLNTWRITE

CLNTWRITE( NumericType Handler, AnyType Object )

Writes one object Object to a communication channel of a client. Handler contains the index of the communication channel opened via CLNTEXEC. See also CLNTREAD, CLNTCLOSE, and CLNTEXEC.

Example:

h2 = clntexec( "nuldrvs -s-" ); . . CLNTWRITE( h2, Model ); . . clntclose( h2,TRUE );

writes the object named Model to client through communication channel h2.

COLOR

COLOR( GeometricType Object, NumericType Color )

Sets the color of the object to one of those specified below. Note that an object has a default color (see IRIT.CFG file) according to its origin - loaded with the LOAD command, PRIMITIVE, or BOOLEAN operation result. The system internally supports colors (although you may have a B&W system) and the colors recognized are: BLACK, BLUE, GREEN, CYAN, RED, MAGENTA, YELLOW, and WHITE.

See the ATTRIB command for more fine control of colors using the RGB attribute. See also AWIDTH and AWIDTH.

This function is equivament to using,

ATTRIB( Object, "color", Width );

COMMENT

COMMENT

Two types of comments are allowed:

1. One-line comment: starts anywhere in a line at the '#' character, up to the end of the line.

2. Block comment: starts at the COMMENT keyword followed by a unique character (anything but white space), up to the second occurrence of that character. This is a fast way to comment out large blocks.

Example:

COMMENT This is a comment

ERROR

ERROR( StringType Message);

Breaks the execution and returns to IRIT main loop, after printing Message to the screen. May be useful in user defined function to break execution in cases of fatal errors.

EXIT

EXIT();

Exits from the solid modeler. NO warning is given!

FOR

FOR( NumericType Start, NumericType Increment, NumericType End, AnyType Body )

Executes the Body (see below), while the FOR loop conditions hold. Start, Increment, End are evaluated first, and the loop is executed while <= End if Increment > 0, or while >= End if Increment < 0. If Start is of the form "Variable = Expression", then that variable is updated on each iteration, and can be used within the body. The body may consist of any number of regular commands, separated by COLONs, including nesting FOR loops to an arbitrary level.

Example:

step = 10; rotstepx = rotx(step); FOR ( a = 1, 1, 360 / step, view_mat = rotstepx * view_mat: view( list( view_mat, axes ), ON ) );

Displays axes with a view direction that is rotated 10 degrees at a time around the X axis.

HELP

HELP( StringType Subject )

Provides help on the specified Subject.

Example:

HELP("");

will list all IRIT help subjects.

FREE

FREE( GeometricType Object )

Because of the usually huge size of geometric objects, this procedure may be used to free them. Reassigning a value (even of different type) to a variable automatically releases the old variable's allocated space as well.

FUNCTION

FuncName = FUNCTION(Prm1, Prm2, ... , PrmN):LclVal1:LclVar2: ... :LclVarM: FuncBody;

Defines a function named FuncName with N parameters and M local variables (N, M >= 0). Here is a (simple) example of a function with no local variables and a single parameter that computes the square of a number:

sqr = FUNCTION(x): return = x * x;

Functions can be defined with optional parameters and optional local variables. A function's body may contain an arbitrary set of expressions including for/while loops, (user) function calls, or even recursive function calls, all separated by colons. The returned value of the function is the value of an automatically defined local variable named return. The return variable is a regular local variable within the scope of the function and can be used as any other variable.

If a variable's name is found in neither the local variable list nor the parameter list, it is searched in the global variable list (outside the scope of the function). Binding of names of variables is static as in the C programming language.

Because binding of variables is performed in execution time, there is a somewhat less restrictive type checking of parameters of functions that are invoked within a user's defined function.

A function can invoke itself, i.e., it can be recursive. However, since a function should be defined when it is called, a dummy function should be defined before the recursive one is defined:

factorial = function(x):return = x; # Dummy function. factorial = function(x): if (x <= 1, return = 1, return = x * factorial(x - 1));

Overloading is valid inside a function as it is outside. For example, for

add = FUNCTION(x, y): return = x + y;

the following function calls are all valid:

add(1, 2); add(vector(1,2,3), point(1,2,3)); add(box(vector(-3, -2, -1), 6, 4, 2), box(vector(-4, -3, -2), 2, 2, 4));

Finally, here is a more interesting example that computes an approximation of the length of a curve, using the sqr function defined above:

distptpt = FUNCTION(pt1, pt2): return = sqrt(sqr(coord(pt1, 1) - coord(pt2, 1)) + sqr(coord(pt1, 2) - coord(pt2, 2)) + sqr(coord(pt1, 3) - coord(pt2, 3))); crvlength = FUNCTION(crv, n):pd:t:t1:t2:dt:pt1:pt2:i: return = 0.0: pd = pdomain(crv): t1 = nth(pd, 1): t2 = nth(pd, 2): dt = (t2 - t1) / n: pt1 = coerce(ceval(crv, t1), e3): for (i = 1, 1, n, pt2 = coerce(ceval(crv, t1 + dt * i), e3): return = return + distptpt(pt1, pt2): pt1 = pt2);

Try, for example:

crvlength(circle(vector(0.0, 0.0, 0.0), 1.0), 30) / 2; crvlength(circle(vector(0.0, 0.0, 0.0), 1.0), 100) / 2; crvlength(circle(vector(0.0, 0.0, 0.0), 1.0), 300) / 2;

See PROCEDURE and IRITSTATE's "DebugFunc" for more.

IF

IF( NumericType Cond, AnyType TrueBody { , AnyType FalseBody } )

Executes TrueBody (group of regular commands, separated by COLONs - see FOR loop) if the Cond holds, i.e., it is a numeric value other than zero, or optionally, if it exists, executes FalseBody if the Cond does not hold, i.e., it evaluates to a numeric value equal to zero.

Examples:

IF ( machine == IBMOS2, resolution = 5, resolution = 10 ); IF ( a > b, max = a, max = b );

sets the resolution to be 10, unless running on an IBMOS2 system, in which case the RESOLUTION variable will be set to 5 in the first statement, and set max to the maximum of a and b in the second statement.

INCLUDE

INCLUDE( StringType FileName )

Executes the script file FileName. Nesting of include file is allowed up to 10 levels deep. If an error occurs, all open files in all nested files are closed and data are waited for at the top level (standard input).

A script file can contain any command the solid modeler supports.

Example:

INCLUDE( "general.irt" );

includes the file "general.irt".

IRITSTATE

IRITSTATE( StringType State, AnyType Data )

Sets a state variable in the IRIT solid modeller. Current supported state variables are,

State Name	Data Type	Comments
Coplanar	NumericType	If TRUE, Coplanar polygons are handled by
		Boolean
DebugMalloc	StringType	If "Reset", memory allocation is cleared/reset.
		No "Free unallocated pointer" test after
		"Reset". If "Print", all allocated blocks are
		printed. Otherwise, used as "address, n": ptr
		address to search for with abort() called after
		n mallocs.
DebugFunc	NumericType	>0 user func. debug information. >2 print params
		on entry, ret. val. on exit. >4 global var. list
		operations.
DumpLevel	NumericType	Controls the way variables/expressions are
		dumped. Only object names/types if >= 0, Scalars
		and vectors are dumped if >= 1, Curves and
		Surfaces are dumped if DumpLvl >= 2,
		Polygons/lines are dumped if DumpLvl >= 3, and
		List objects are traversed recursively if
		DumpLvl >= 4.
EchoSource	NumericType	If TRUE, irit scripts are echoed to stdout.
FloatFrmt	StringType	Specifies a new printf floating point format.
InterCrv	NumericType	If TRUE Boolean operations creates only
		intersection curves. If FALSE, full Boolean
		operation results.
InterpProd	ConstantType	TRUE for Bspline sym. products via
		interpolation FALSE for Bspline sym. products
		via Bezier.
PolySort	NumericType	Axis of Polygon Intersection sweep in Boolean
		operations: 0 for X axis, 1 for Y axis, 2 for
		Z axis.
PrimRatSrfs	NumericType	TRUE for rational exact primitive surfaces,
		FALSE for approximated polynomial (integral)
		surfaces. See also PrimSrfs.
PrimSrfs	NumericType	TRUE for primitive construction as freeform
		surfaces, FALSE for polygonal primitives.
		See also PrimRatSrfs.
TrimCrvs	NumericType	Number of samples the higher order trimmed
		curves are sampled, in piecewise linear
		approximation. If zero, computed
		symbolically as composition.
UVBoolean	NumericType	If TRUE, Boolean between surfaces returns UV
		instead of Euclidean curves.

Example:

IRITSTATE( "DebugFunc", 3 ); IRITSTATE( "FloatFrmt", "%8.5lg" );

To print parameters of user defined functions on entry, and return value on exit. Also selects a floating point printf format of "%8.5lg".

INTERACT

INTERACT( GeometryTreeType Object )

A user-defined function (see iritinit.irt) that does the following, in order:

Clear the display device.

Display the given Object.

Pause for a keystroke.

This user-defined function in version 4.0 of IRIT is an emulation of the INTERACT function that used to exist in previous versions.

Example:

INTERACT( list( view_mat, Axes, Obj ) );

displays and interacts with the object Obj and the predefined object Axes. VIEW_MAT will be used to set the starting transformation.

See VIEW and VIEWOBJ for more.

LIST

ListType LIST( AnyType Elem1, AnyType Elem2, ... )

Constructs an object as a list of several other objects. Only a reference is made to the Elements, so modifying Elem1 after being included in the list will affect Elem1 in that list next time list is used!

Each inclusion of an object in a list increases its internal used reference. The object is freed iff in used reference is zero. As a result, attempt to delete a variable (using FREE) which is referenced in a list removes the variable, but the object itself is freed only when the list is freed.

LOAD

AnyType LOAD( StringType FileName )

Loads an object from the given FileName. The object may be any object defined in the system, including lists, in which the structure is recovered and reconstructed as well (internal objects are inserted into the global system object list if they have names). If no file type is provided, ".dat" is assumed.

This command can also be used to load binary files. Ascii regular data files are usually loaded in much more time then binary files due the the parsing required. Binary data files can be loaded directly like ascii files in IRIT, but can only be inspected through IRIT tools such as dat2irit. A binary data file must have a ".bdt" (Binary DaTa) type in its name.

Under unix, compressed files can be loaded if the given file name has a postfix of ".Z". The unix system's "zcat" will be invoked via a pipe for that purpose.

LOGFILE

LOGFILE( NumericType Set ) or LOGFILE( StringType FileName )

If Set is non zero (see TRUE/FALSE and ON/OFF), then everything printed in the input window, will go to the log file specified in the IRIT.CFG configuration file. This file will be created the first time logfile is turned ON. If a string FileName is provided, it will be used as a log file name from now on. It also closes the current log file. A "LOGFILE( on );" must be issued after a log file name change.

Example:

LOGFILE( "Data1" ); LOGFILE( on ); printf( "Resolution = %lf\n", list( resolution ) ); LOGFILE( off );

to print the current resolution level into file Data1.

MSLEEP

MSLEEP( NumericType MilliSeconds )

Causes the solid modeller to sleep for the prescribed time in milliseconds.

Example:

for ( i = 1, 1, sizeof( crvs ), c = nth( crvs, i ): color( c, yellow ): msleep(20): viewobj( c ) );

Displays an animation sequence and sleeps for 20 milliseconds between iterations.

NTH

AnyType NTH( ListType ListObject, NumericType Index )

Returns the Index (base count 1) element of the list ListObject.

Example:

Lst = list( a, list( b, c ), d ); Lst2 = NTH( Lst, 2 );

and now Lst2 is equal to 'list( b, c )'.

PAUSE

PAUSE( NumericType Flush )

Waits for a keystroke. Nice to have if a temporary stop in a middle of an included file (see INCLUDE) is required. If Flush is TRUE, then the input is first flushed to guarantee that the actual stop will occur.

PRINTF

PRINTF( StringType CtrlStr, ListType Data )

A formatted printing routine, following the concepts of the C programming language's printf routine. CtrlStr is a string object for which the following special '%' commands are supported:

%d, %i, %u	Prints the numeric object as an integer or unsigned integer.
%o, %x, %X	Prints the numeric object as an octal or hexadecimal integer.
%e, %f, %g,	Prints the numeric object in several formats of
%E, %F	floating point numbers.
%s	Prints the string object as a string.
%pe, %pf, %pg	Prints the three coordinates of the point object.
%ve, %vf, %vg	Prints the three coordinates of the vector object.
%Pe, %Pf, %Pg,	Prints the four coordinates of the plane object.
%De, %Df, %Dg,	Prints the given object in IRIT's data file format.

All the '%' commands can include any modifier that is valid in the C programming language printf routine, including l (long), prefix character(s), size, etc. The point, vector, plane, and object commands can also be modified in a similar way, to set the format of the numeric data printed.

Also supported are the newline and tab using the backslash escape character:

* PRINTF("\tThis is the char "\%"\n", nil());

Backslashes should be escaped themselves as can be seen in the above example. Here are few more examples:

* PRINTF("this is a string "%s" and this is an integer %8d.\n", * list("STRING", 1987)); * PRINTF("this is a vector [%8.5lvf]\n", list(vector(1,2,3))); * IritState("DumpLevel", 9); * PRINTF("this is a object %8.6lDf...\n", list(axes)); * PRINTF("this is a object %10.8lDg...\n", list(axes));

This implementation of PRINTF is somewhat different than the C programming language's version, because the backslash always escapes the next character during the processing stage of IRIT's parser. That is, the string

* '\tThis is the char "\%"\n'

is actually parsed by the IRIT's parser into

* 'tThis is the char "%"n'

because this is the way the IRIT parser processes strings. The latter string is the one that PRINTF actually see.

PROCEDURE

ProcName = PROCEDURE(Prm1, Prm2, ... , PrmN):LclVal1:LclVar2: ... :LclVarM: ProcBody;

A procedure is a function that does not return a value, and therefore the return variable (see FUNCTION) should not be used. A procedure is identical to a function in every other way. See FUNCTION for more.

RMATTR

RMATTR( AnyType Object, StringType Name )

Removes attribute named Name from object Object. This function will have no affect on Object if Object have no attribute named Name.

See also ATTRIB.

SAVE

SAVE( StringType FileName, AnyType Object )

Saves the provided Object in the specified file name FileName. No extension type is needed (ignored if specified), and ".dat" is supplied by default. Object can be any object type, including list, in which structure is saved recursively. See also LOAD. If a display device is actively running at the time SAVE is invoked, its transformation matrix will be saved with the same name but with extension type of ".mat" instead of ".dat".

This command can also be used to save binary files. Ascii regular data files are usually loaded in much more time then binary files due the the parsing required. Binary data files can be loaded directly like ascii files in IRIT, but must be inspected through IRIT tools such as dat2irit. A binary data file must have a ".bdt" (Binary DaTa) type in its name.

Under unix, files will be saved compressed if the given file name has a postfix of ".Z". The unix system's "compress" will be invoked via a pipe for that purpose.

Example:

SAVE( "Obj1.bdt.Z", Obj1 );

Saves Obj1 in the file Obj1.bdt.Z as compressed binary file.

SNOC

SNOC( AnyType Object, ListType ListObject )

Similar to the lisp cons operator but puts the new Object in the end of the list ListObject instead of the beginning, in place.

Example:

Lst = list( axes ); SNOC( Srf, Lst );

and now Lst is equal to the list 'list( axes, Srf )'.

SYSTEM

SYSTEM( StringType Command )

Executes a system command Command. For example,

SYSTEM( "ls -l" );

TIME

TIME( NumericType Reset )

Returns the time in seconds from the last time TIME was called with Reset TRUE. This time is CPU time if such support is available from the system (times function), and is real time otherwise (time function). The time is automatically reset at the beginning of the execution of this program.

Example:

Dummy = TIME( TRUE ); . . . TIME( FALSE );

prints the time in seconds between the above two time function calls.

VARLIST

VARLIST()

List all the currently defined objects in the system.

VECTOR

VectorType VECTOR( NumericType X, NumericType Y, NumericType Z )

Creates a vector type object, using the three provided NumericType scalars.

VIEW

VIEW( GeometricTreeType Object, NumericType ClearWindow )

Displays the (geometric) object(s) as given in Object.

If ClearWindow is non-zero (see TRUE/FALSE and ON/OFF) the window is first cleared (before drawing the objects).

Example:

VIEW( Axes, FALSE );

displays the predefined object Axes in the viewing window on top of what is drawn already.

In version 4.0, this function is emulated (see iritinit.irt) using the VIEWOBJ function. In order to use the current viewing matrix, VIEW_MAT should be provided as an additional parameter. For example,

VIEW( list( view_mat, Obj ), TRUE );

However, since VIEW is a user defined function, the following will not use VIEW_MAT as one would expect:

VIEW( view_mat, TRUE );

because VIEW_MAT will be renamed inside the VIEW user defined function to a local (to the user defined function) variable.

In iritinit.irt one can find several other useful VIEW related functions:

VIEWCLEAR	Clears all data displayed on the display device.
VIEWREMOVE	Removes the object specified by name from display.
VIEWDISC	Disconnects from display device (which is still running)
	while allowing IRIT to connect to a new device.
VIEWEXIT	Forces the display device to exit.
VIEWSAVE	Request sdisplay device to save transformation matrix.
BEEP	An emulation of the BEEP command of versions prior to 4.0.
VIEWSTATE	Allows to change the state of the display device.

For the above VIEW related functions, only VIEWREMOVE, VIEWSAVE, and VIEWSTATE require a parameter, which is the file name and view state respectively. The view state can be one of several commands. See the display device section for more.

Examples:

VIEWCLEAR(); VIEW( axes, off ); VIEWSTATE( "LngrVecs" ); VIEWSTATE( "DrawSolid" ); VIEWSAVE( "matrix1" ); VIEWREMOVE( "axes" ); VIEWDISC();

VIEWOBJ

VIEWOBJ( GeometricTreeType Object )

Displays the (geometric) object(s) as given in Object. Object may be any GeometricType or a list of other GeometricTypes nested to an arbitrary level.

Unlike IRIT versions prior to 4.0, VIEW_MAT is not explicitly used as the transformation matrix. In order to display with a VIEW_MAT view, VIEW_MAT should be listed as an argument (in that exact name) to VIEWOBJ. Same is true for the perspective matrix PRSP_MAT.

Example:

VIEWOBJ( list( view_mat, Axes ) );

displays the predefined object Axes in the viewing window using the viewing matrix VIEW_MAT.

WHILE

WHILE( NumericType Cond, AnyType Body )

Executes the Body (see below), while the WHILE loop conditions Cond is evaluated into a non zero value. Cond is being evaluated before each iteration.

The body may consist of any number of regular commands, separated by COLONs, including nesting loops to an arbitrary level.

Example:

deg = 0; rotstepx = rotx( 10 ); WHILE ( deg < 360, deg = deg + 10: view_mat = rotstepx * view_mat: view( list( view_mat, axes ), ON ) );

Displays axes with a view direction that is rotated 10 degrees at a time around the X axis.

System variables

System variables are predefined objects in the system. Any time IRIT is executed, these variable are automatically defined and set to values which are sometimes machine dependent. These are regular objects in any other sense, including the ability to delete or overwrite them. One can modify, delete, or introduce other objects using the IRITINIT.IRT file.

AXES

Predefined polyline object (PolylineType) that describes the XYZ axes.

DRAWCTLPT

Predefined Boolean variable (NumericType) that controls whether curves' control polygons and surfaces' control meshes are drawn (TRUE) or not (FALSE). Default is FALSE.

FLAT4PLY

Predefined Boolean object (NumericType) that controls the way almost flat surface patches are converted to polygons: four polygons (TRUE) or only two polygons (FALSE). Default value is FALSE.

MACHINE

Predefined numeric object (NumericType) holding the machine type as one of the following constants: MSDOS, SGI, HP, SUN, APOLLO, UNIX, IBMOS2, IBMNT, and AMIGA.

POLY_APPROX_OPT

A variable controlling the algorithm to convert surfaces to polygons. This two digit number controls the method that is used to subdivide a surface into polygons. The first digit (units) can be one of:

0	Uniform sampling in a fixed grid.
1	An alternated U and V subdivision direction. Once U is
	subdivided and then V is subdivided.
2	A min max subdivision direction. In other words, the
	direction that minimizes the maximal error is selected.

The second digit (tenths) can be one of:

0	A fixed sized regular grid. The side of the grid is set
	via the RESOLUTION variable.
1	This mode is not for general use.
2	Maximal distance between the surface and its polygonal
	approximation is bounded by bilinear surface fit.
	Maximal distance allowed is set via POLY_APPROX_TOL.
	Recommended choice for optimal polygonization.
3	This mode is not for general use.

POLY_APPROX_UV

A Boolean predefined variable. If TRUE, UV values of surface polygonal approximation are placed on attribute lists of vertices.

POLY_APPROX_TOL

A numeric predefined tolerance control on the distance between the surface and its polygonal approximation in POLY_APPROX_OPT settings.

PRSP_MAT

Predefined matrix object (MatrixType) to hold the perspective matrix used/set by VIEW and/or INTERACT commands. See also VIEW_MAT.

RESOLUTION

Predefined numeric object (NumericType) that sets the accuracy of the polygonal primitive geometric objects and the approximation of curves and surfaces. Holds the number of divisions a circle is divided into (with minimum value of 4). If, for example, RESOLUTION is set to 6, then a generated CONE will effectively be a six-sided pyramid. Also controls the fineness of freeform curves and surfaces when they are approximated as piecewise linear polylines, and the fineness of freeform surfaces when they are approximated as polygons.

VIEW_MAT

Predefined matrix object (MatrixType) to hold the viewing matrix used/set by VIEW and/or INTERACT commands. See also PRSP_MAT.

System constants

The following constants are used by the various functions of the system to signal certain conditions. Internally, they are represented numerically, although, in general, their exact value is unimportant and may be changed in future versions. In the rare circumstance that you need to know their values, simply type the constant as an expression.

Example:

MAGENTA;

AMIGA

A constant designating an AMIGA system, in the MACHINE variable.

APOLLO

A constant designating an APOLLO system, in the MACHINE variable.

BLACK

A constant defining a BLACK color.

BLUE

A constant defining a BLUE color.

COL

A constant defining the COLumn or U direction of a surface or a trivariate mesh.

CTLPT_TYPE

A constant defining an object of type control point.

CURVE_TYPE

A constant defining an object of type curve.

CYAN

A constant defining a CYAN color.

DEPTH

A constant defining the DEPTH direction of a trivariate mesh. See TBEZIER, TBSPLINE.

E1

A constant defining an E1 (X only coordinate) control point type.

E2

A constant defining an E2 (X and Y coordinates) control point type.

E3

A constant defining an E3 (X, Y and Z coordinates) control point type.

E4

A constant defining an E4 control point type.

E5

A constant defining an E5 control point type.

FALSE

A zero constant. May be used as Boolean operand.

GREEN

A constant defining a GREEN color.

HP

A constant designating an HP system, in the MACHINE variable.

IBMOS2

A constant designating an IBM system running under OS2, in the MACHINE variable.

IBMNT

A constant designating an IBM system running under Windows NT, in the MACHINE variable.

KV_FLOAT

A constant defining a floating end condition uniformly spaced knot vector.

KV_OPEN

A constant defining an open end condition uniformly spaced knot vector.

KV_PERIODIC

A constant defining a periodic end condition with uniformly spaced knot vector.

LIST_TYPE

A constant defining an object of type list.

MAGENTA

A constant defining a MAGENTA color.

MATRIX_TYPE

A constant defining an object of type matrix.

MSDOS

A constant designating an MSDOS system, in the MACHINE variable.

NUMERIC_TYPE

A constant defining an object of type numeric.

OFF

Synonym of FALSE.

ON

Synonym for TRUE.

P1

A constant defining a P1 (W and WX coordinates, in that order) rational control point type.

P2

A constant defining a P2 (W, WX, and WY coordinates, in that order) rational control point type.

P3

A constant defining a P3 (W, WX, WY, and WZ coordinates, in that order) rational control point type.

P4

A constant defining a P4 rational control point type.

P5

A constant defining a P5 rational control point type.

PARAM_CENTRIP

A constant defining a centripetal length parametrization.

PARAM_CHORD

A constant defining a chord length parametrization.

PARAM_UNIFORM

A constant defining an uniform parametrization.

PI

The constant of 3.141592...

PLANE_TYPE

A constant defining an object of type plane.

POINT_TYPE

A constant defining an object of type point.

POLY_TYPE

A constant defining an object of type poly.

RED

A constant defining a RED color.

ROW

A constant defining the ROW or V direction of a surface or a trivariate mesh.

SGI

A constant designating an SGI system, in the MACHINE variable.

STRING_TYPE

A constant defining an object of type string.

SURFACE_TYPE

A constant defining an object of type surface.

SUN

A constant designating a SUN system, in the MACHINE variable.

TRIMSRF_TYPE

A constant defining an object of type trimmed surface.

TRIVAR_TYPE

A constant defining an object of type trivariate function.

TRUE

A non zero constant. May be used as Boolean operand.

UNDEF_TYPE

A constant defining an object of no type (yet).

UNIX

A constant designating a generic UNIX system, in the MACHINE variable.

VECTOR_TYPE

A constant defining an object of type vector.

WHITE

A constant defining a WHITE color.

YELLOW

A constant defining a YELLOW color.

Animation

The animation tool adds the capability of animating objects using forward kinematics, exploiting animation curves. Each object has different attributes, that prescribe its motion, scale, and visibility as a function of time. Every attribute has a name, which designates it's role. For instance an attribute animation curve named MOV_X describes a translation motion along the X axis.

How to create animation curves in IRIT

Let OBJ be an object in IRIT to animate.

Animation curves are either scalar (E1/P1) curves or three dimensional (E3/P3) curves with one of the following name prefixes:

MOV_X, MOV_Y, MOV_Z	Translation along one axis
MOV_XYZ	Arbitrary translation along all three axes
ROT_X, ROT_Y, ROT_Z	Rotating around a single axis (degrees)
SCL_X, SCL_Y, SCL_Z	Scale along a single axis
SCL	Global scale
VISIBLE	Visibility

The visibility curve is a scalar curve that enables the display of the object if the visibility curve is positive at time t and disables the display (hide) the object if the visibility curve is negative at time t.

The animation curves are all attached as an attribute named "animation" to the object OBJ.

Example:

mov_x = cbezier( ctlpt( E1, 0.0 ), ctlpt( E1, 1.0 ) ); scl = cbezier( ctlpt( E1, 1.0 ), ctlpt( E1, 0.1 ) ); rot_y = cbezier( ctlpt( E1, 0.0 ), ctlpt( E1, 0.0 ) ); ctlpt( E1, 360.0 ) ); attrib(OBJ, "animation", list( mov_x, scl, rot_y ) );

To animate OBJ between time zero and one (Bezier curves are always between zero and one), by moving it a unit size in the X direction, scaling it to %10 of its original size and rotating it at increasing angular speed from zero to 360 degrees.

OBJ can now be save into a file or displayed via one of the regular viewing commands in IRIT (i.e. VIEWOBJ).

Animation is not always between zero and one. To that end, one can apply the CREPARAM function to modify the parametric domain of the animation curve. The convention is that if the time is below the starting value of the parametric domain, the starting value of the curve is used. Similarly if the time is beyond the end of the parameter domain of the animation curve, the end value of the animation curve is used.

Example:

CREPARAM( mov_x, 3.0, 5.0 );

to set the time of the motion in the x axis to be from t = 3 to t = 5. for t < 3, mov_x(3) is used, and for t > 5, mov_x(5) is employed.

the animation curves are regular objects in the IRIT system. Hence, only one object named mov_x or scl can exist at one time. If you create a new object named mov_x, the old one is overwritten! To preserve old animation curves you can detach the old ones by executing 'free(mov_x)' that removes the object named mov_x from IRIT's object list but not from its previous used locations within other list objects, if any. A different way to it call this animation curves mov_x1, mov_x2 etc. as only the prefix of the name is verified.

For example:

mov_x = cbezier( ctlpt( E1, 0.0 ), ctlpt( E1, 1.0 ) ); attrib(obj1, "animation", list( mov_x ) ); free(mov_x); mov_x1 = cbezier( ctlpt( E1, 2.0 ), ctlpt( E1, 3.0 ) ); mov_x2 = cbezier( ctlpt( E1, 2.0 ), ctlpt( E1, 3.0 ) ); attrib(obj2, "animation", list( mov_x1, mov_x2 ) ); free(mov_x);

notice the way we have two animation curves translating obj2 in x. This is somewhat artificial but might make more sense if other transformations would appear in between.

A more complete animation example

a = box( vector( 0, 0, 0 ), 1, 1, 1 ); b = box( vector( 0, 0, 0 ), 1, 1, 1 ); c = box( vector( 0, 0, 0 ), 1, 1, 1 ); d = sphere( vector( 0, 0, 0), 0.7 ); pt0 = ctlpt( e1, 0.0 ); pt1 = ctlpt( e1, 1.0 ); pt2 = ctlpt( e1, 2.0 ); pt6 = ctlpt( e1, 6.0 ); pt360 = ctlpt( e1, 360.0 ); pt10 = ctlpt( e1, -4.0 ); pt11 = ctlpt( e1, 1.0 ); pt12 = ctlpt( e1, 4.0 ); pt13 = ctlpt( e1, -1.0 ); visible = creparam( cbezier( list( pt10, pt11 ) ), 0.0, 5.0 ); mov_x = creparam( cbezier( list( pt0, pt6, pt2 ) ), 0.0, 1.2 ); mov_y = mov_x; mov_z = mov_x; rot_x = creparam( cbspline( 2, list( pt0, pt360, pt0 ), list( KV_OPEN ) ), 1.2, 2.5 ); rot_y = rot_x; rot_z = rot_x; scl = creparam( cbezier( list( pt1, pt2, pt1, pt2, pt1 ) ), 2.5, 4.0 ); scl_x = scl; scl_y = scl; scl_z = scl; mov_xyz = creparam( circle( vector( 0, 0, 0 ), 2.0 ), 4.0, 5.0 ); attrib( d, "animation", list( mov_xyz, visible ) ); free( visible ); visible = creparam( cbezier( list( pt12, pt13 ) ), 0.0, 5.0 ); attrib( a, "animation", list( rot_x, mov_x, scl, scl_x, visible ) ); attrib( b, "animation", list( rot_y, mov_y, scl, scl_y, visible ) ); attrib( c, "animation", list( rot_z, mov_z, scl, scl_z, visible ) ); color( a, red ); color( b, green ); color( c, blue ); color( d, cyan ); demo = list( a, b, c, d ); interact( demo ); viewanim( 0, 5, 0.01 );

In this example, we create four objects, three cubes and one sphere. Animation curves to translate the three cubes along the three axes for the time period of t = 0 to t = 1.2 are created. Rotation curves to rotate the three cubes along the three axes are then created for time period of t = 1.2 to t = 2.5. Finally, for the time period of t = 2.5 to t = 4.0. the cubes are (not only) unifomly scaled. For the time period of t = 4 to t = 5, the cubes become invisible and the sphere, that becomes visible, is rotated along a circle of radius 2.

Another complete animation example

This example demonstrates the ability to put "animation" attributes on internal nodes of a hierarchy, affecting the entire set of objects in the hierachy. Herein, we present an robotic arm with three edges and two joints.

BoxLength = 2; BoxWidth = 2; BoxHeight = 10; LowerBox = box( vector( -BoxLength / 2, -BoxWidth / 2, 0 ), BoxLength, BoxWidth, BoxHeight); MiddleBox = box( vector( -BoxLength / 2, -BoxWidth / 2, 0 ), BoxLength, BoxWidth, BoxHeight); UpperBox = box( vector( -BoxLength / 2, -BoxWidth / 2, 0 ), BoxLength, BoxWidth, BoxHeight); Cn1 = cone( vector( 0, 0, 0 ), vector( 0, BoxHeight / 3, 0 ), 1 ); color( LowerBox, magenta ); color( MiddleBox, yellow ); color( UpperBox, cyan ); color( Cn1, green ); rot_x1 = creparam( cbspline( 3, list( ctlpt( E1, 0 ), ctlpt( E1, -200 ), ctlpt( E1, 200 ), ctlpt( E1, 0 ) ), list( KV_OPEN ) ), 0, 3 ); rot_x2 = creparam( cbspline( 4, list( ctlpt( E1, 0 ), ctlpt( E1, 400 ), ctlpt( E1, -400 ), ctlpt( E1, 0 ) ), list( KV_OPEN ) ), 0, 3 ); rot_y = creparam( cbspline( 2, list( ctlpt( E1, 0 ), ctlpt( E1, 100 ), ctlpt( E1, -100 ), ctlpt( E1, 0 ) ), list( KV_OPEN ) ), 0, 3 ); rot_z = creparam( cbspline( 2, list( ctlpt( E1, 0 ), ctlpt( E1, 1440 ) ), list( KV_OPEN ) ), 0, 3 ); Translate = trans( vector( 0, 0, BoxHeight ) ); attrib( Cn1, "animation", list( rot_z, Translate ) ); Upr = list( Cn1, UpperBox ); attrib( Upr, "animation", list( rot_y, Translate ) ); Mid = list( Upr, MiddleBox ); attrib( Mid, "animation", list( rot_x2, Translate ) ); rbt_hand = list( Mid, LowerBox ); attrib( rbt_hand, "animation", list( rot_x1 ) ); view( rbt_hand, 1 );

In this example, we create four objects, three cubes and one cone, simulating a robotic hand with three edges an a gripper (the cone). The animation is define hierarchically, making it very easy to model the robot.

Display devices

The following display device drivers are available,

Device Name	Invocation	Environment

xgldrvs	xgldrvs -s-	SGI 4D GL regular driver.
xogldrvs	xogldrvs -s-	SGI 4D Open GL/Motif driver.
xgladap	xgladap -s-	SGI 4D GL adaptive isocurve
		experimental driver.
x11drvs	x11drvs -s-	X11 driver.
xmtdrvs	xmtdrvs -s-	X11 Motif driver.
xglmdrvs	xglmdrvs -s-	SGI 4D GL and X11/Motif driver.
wntdrvs	wntdrvs -s-	IBM PC Windows NT driver.
wntgdrvs	wntgdrvs -s-	IBM PC Windows NT Open GL driver.
os2drvs	os2drvs -s-	IBM PC OS2 2.x/3.x driver.
amidrvs	amidrvs -s-	AmigaDOS 2.04+ driver.
nuldrvs	nuldrvs -s- [-d] [-D]	A device to print the
		object stream to stdout.

All display devices are clients communicating with the server (IRIT) using IPC (inter process communication). On Unix and Window NT sockets are used. A Windows NT client can talk to a server (IRIT) on a unix host if hooked to the same netwrok. On OS2 pipes are used, and both the client and server must run on the same machine. On AmigaDOS exec messages are used, and both the client and server must run on the same machine.

While all display devices support object(s) transformations via a transformation control window, many of the display devices allow one to click and drag on the viewing window to rotate (Left Button) and to translate (Right Button). This mode exploits the two degrees of freedom of the mouse to provide intuitive dual axis rotation and translation.

The server (IRIT) will automatically start a client display device if the IRIT_DISPLAY environment variable is set to the name and options of the display device to run. For example:

setenv IRIT_DISPLAY xgldrvs -s-

The display device must be in a directory that is in the path environment variable. Most display devices require the '-s-' flags to run in a non-standalone mode, or a client-server mode. Most drivers can also be used to display data in a standalone mode (i.e., no server). For example:

xgldrvs -s solid1.dat irit.mat

Effectively, all the display devices are also data display programs (poly3d, which was the display program prior to version 4.0, is retired in 4.0). Therefore some functionality is not always as expected. For example, the quit button will always force the display device to quit, even if poped up from IRIT, but will not cause IRIT to quit as might logically expected. In fact, the next time IRIT will try to communicate with the display device, it will find the broken connection and will start up a new display device.

Most display devices recognize attributes found on objects. The following attributes are usually recognized (depending on the device capability.):

COLOR: Selects the drawn color of the object to be one of the 8/16 predefined colors in the IRIT system: white, red, green, blue, yellow, cyan, magenta, black.

RGB: Overwrites (if supported) the COLOR attribute (if given) and sets the color of the object to the exact prescribed RGB set.

DWIDTH: Sets the width in pixels of the drawn object, when drawn as a wireframe.

All display devices recognize all the command line flags and all the configuration options in a configuration file, as described below. The display devices will make attemps to honor the requests, to the best of their ability. For example, only xgldrvs can render shaded models, and so only it will honor a DrawSolid configuration options.

Command Line Options

???drvs [-s] [-u] [-n] [-N] [-i] [-c] [-C] [-m] [-a] [-g x1,x2,y1,y2] [-G x1,x2,y1,y2] [-I #IsoLines] [-F PolygonOpti FineNess] [-f PolylineOpti SampTol] [-l LineWidth] [-r] [-A Shader] [-B] [-2] [-d] [-D] [-L NormalLen] [-4] [-b BackGround] [-S LgtSrcPos] [-Z ZMin ZMax] [-M] [-P] [-x ExecAnimCmd] [-X Min,Max,Dt{,flags}] [-z] DFiles

-s: Runs the driver in a Standalone mode. Otherwise, the driver will attempt to communicate with the IRIT server.

-u: Forces a Unit matrix. That is, input data are not transformed at all.

-n: Draws normals of vertices.

-N: Draws normals of polygons.

-i: Draws internal edges (created by IRIT) - default is not to display them, and this option will force displaying them as well.

-c: Sets depth cueing on. Drawings that are closer to the viewer will be drawn in more intense color.

-C: Cache the piecewise linear geometry so curves and surface can be redisplay faster. Purging it will free memory, on the other hand.

-m: Provides some more information on the data file(s) parsed.

-a: Sets the support of antialiased lines.

-g x1,x2,y1,y2: Prescribes the position and location of the transformation window by prescribing the domain of the window in screen space pixels.

-G x1,x2,y1,y2: Prescribes the position and location of the viewing window by prescribing the domain of the window in screen space pixels.

-I #IsoLines: Specifies number of isolines per surface, per direction. A specification of zero isolines is possible only on the command line and it denotes the obvious.

-F PolyOpti FineNess: Controls the method used to approximate surfaces into polygons. See the variable POLY_APPROX_OPT for the meaning of FineNess. See also -4.

-f PolyOpti SampTol: Controls the method used to approximate curves into polylines. If PolyOpti == 0, equally spaced intervals are used. For PolyOpti == 1, an adaptive subdivision that optimizes the samples is employed. For PolyOpti == 2, SampTol (real number) specifies the maximal allowed deviation tolerance of the piecewise linear approximation from the original curve. Default is 0 64 (uniform sampling with 64 samples).

-l LineWidth: Sets the linewidth, in pixels. Default is one pixel wide.

-r: Rendered mode. Draws object as solid.

-A Shader: Shader can be one of 1 (Flat), 2 (Gouraud), or 3 (Phong).

-B: Back face culling of polygons.

-2: Double buffering. Prevents screen flicker on the expense of possibly less colors.

-d: Debug objects. Prints to stderr all objects read from communcation port with the server IRIT.

-D: Debug input. Prints to stderr all characters read from communcation port with the server IRIT. Lowest level of communication.

-L NormalLen: Sets the length of the drawn normals in thousandths of a unit.

-4: Forces four polygons per almost flat region in the surface to polygon conversion. Otherwise two polygons only.

-b BackGround: Sets the background color as three RGB integers in the range of 0 to 255.

-S LgtSrcPos: Sets the lighting via the light source position.

-Z ZMin ZMax: Sets the near and far Z clipping planes.

-M: Draw control mesh/polygon of curves and surfaces, as well.

-x ExecAnimCmd: Command to execute as a subprocess every iteration of display of an animation sequence. This command can for example save the display into an image file, saving the animation sequence. One parameter is passed, which is an running index starting from one.

-X Min,Max,Dt{,flags}: Executes an animation sequence between Min time to Max time is steps of Dt. If an 's' flag is specified the animation is saved as idividual data files, one per frame, for high quality rendering. If 'x' flag is given, the display device exists once the animation is complete.

-P: Draws curves and surfaces (surfaces are drawn using a set of isocurves, see -I, or polygons, see -f).

-z: Prints version number and current defaults.

Configuration Options

The configuration file is read before the command line options are processed. Therefore, all options in this section can be overriden by the appropriate command line option, if any.

TransPrefPos: Preferred location (Xmin, YMin, Xmax, Ymax) of the transformation window.

ViewPrefPos: Preferred location (Xmin, YMin, Xmax, Ymax) of the viewing window.

BackGround: Background color. Same as '-b'.

Internal: Draws internal edges. Same as '-i'.

LightSrcPos: Sets the location of the (first) light source as a rational four coefficient location. W of zero sets the light source at infinity.

ExecAnimCmd: Executes a command each step of animation. Same as '-x'.

ExecAnimation: Executes an animation sequence on startup. Same as '-X'.

DrawVNormal: Draws normals of vertices. Same as '-n'.

DrawPNormal: Draws normals of polygons. Same as '-n'.

MoreVerbose: Provides some more information on the data file(s) parsed. Same as '-m'.

UnitMatrix: Forces a Unit matrix. That is, input data are not transformed at all. Same as '-u'.

DrawSolid: Requests a shaded surface rendering, as opposed to isocurve surface representation.

BFaceCull: Requests the removal of back facing polygons, for better visibility.

DoubleBuffer: Requests drawing using a double buffer, if any.

DebugObjects: Debug objects. Prints to stderr all objects read from the communcation port with the server IRIT. Same as '-d'.

DebugEchoInput: Debug input. Prints to stderr all characters read from the communcation port with the server IRIT. Lowest level of communications.

DepthCue: Set depth cueing on. Drawings that are closer to the viewer will be drawn in more intense color. Same as '-c'.

CacheGeom: Normally piecewise linear approximation of freefroms is cached. By setting this option to FALSE, no such auxiliary data is being saved, reducing the memory overhead. Same as '-C'.

FourPerFlat: Forces four polygons per almost flat region in the surface to polygon conversion. Otherwise two polygons only. Same as '-4'.

AntiAlias: Request the drawing of anti aliased lines.

DrawSurfaceMesh: Draws control mesh/polygon of curves and surfaces, as well. Same as '-M'.

DrawSurfacePoly: Draws curves and surfaces (surfaces are drawn using a set of isocurves, see -I, or polygons, see -f). Same as '-P'.

StandAlone: Runs the driver in a Stand alone mode. Otherwise, the driver will attempt to communicate with the IRIT server. Same as '-s'.

NumOfIsolines: Specifies number of isolines per surface, per direction. Same as '-I'.

SamplesPerCurve: Specifies the samples per (iso)curve. See '-f'.

LineWidth: Sets the linewidth, in pixels. Default is one pixel wide. Same as '-l'

AdapIsoDir: Selects the direction of the adaptive isoline rendering.

PolygonOpti: Controls the method used to subdivide a surface into polygons that approximate it. Same as '-F'.

PolylineOpti: Controls the method used to subdivide a curve into polylines that approximate it. Same as '-f'.

ShadingModel: One of 1 (Flat), 2 (Gouraud), or 3 (Phong). Same as '-A'.

TransMode: Selects between object space transformations and screen space transformation.

ViewMode: Selects between perspective and orthographic views.

NormalLength: Sets the length of the drawn normals in thousandths of a unit. Same as '-L'.

ZClipMin: Sets the minimal clipping plane in Z. Same as '-Z'.

ZClipMax: Sets the maximal clipping plane in Z. Same as '-Z'.

FineNess: Controls the fineness of the surface to polygon subdivision. See '-F'.

Interactive mode setup

Commands that affect the status of the display device can also be sent via the communication port with the IRIT server. The following commands are recognized as string objects with object name of "COMMAND_":

BEEP	Makes some sound.
CLEAR	Clears the display area. All objects are deleted.
DCLEAR	Delayed clear. Same as CLEAR but delayed until next
	object is sent from the server. Useful for animation.
DISCONNECT	Closes connection with the server, but does not quit.
EXIT	Closes connection with the server and quits.
GETOBJ NAME	Requests the object named NAME that is returned
	in the output channel to the server.
MSAVE NAME	Saves the transformation matrix file by the name
	NAME.
REMOVE NAME	Request the removal of object named NAME from
	display.
ANIMATE TMin TMax Dt	Animates current scene from TMin to TMax in Dt
	steps.
STATE COMMAND	Changes the state of the display device. See below.

The following commands are valid for the STATE COMMAND above,

MoreSense:	More sensitive mouse control.
LessSense:	Less sensitive mouse control.
ScrnObject:	Toggle screen/object transformation mode.
PerspOrtho:	Toggles perspective/orthographic trans. mode.
DepthCue:	Toggles depth cueing drawing.
DrawSolid:	Toggles isocurve/shaded solid drawing.
ShadingMdl:	Toggles shading model for solid solid drawing.
BFaceCull:	Cull back facing polygons.
DblBuffer:	Toggles single/double buffer mode.
AntiAlias:	Toggles anti aliased lines.
DrawIntrnl:	Toggles drawing of internal lines.
DrawVNrml:	Toggles drawing of normals of vertices.
DrawPNrml:	Toggles drawing of normals of polygons.
DSrfMesh:	Toggles drawing of control meshes/polygons.
DSrfPoly:	Toggles drawing of curves/surfaces.
4PerFlat:	Toggles 2/4 polygons per flat surface regions.
MoreIso:	Doubles the number of isolines in a surface.
LessIso:	Halves the number of isolines in a surface.
FinrAprx:	Doubles the number of samples per curve.
CrsrAprx:	Halves the number of samples per curve.
LngrVecs:	Doubles the length of displayed normal vectors.
ShrtrVecs:	Halves the length of displayed normal vectors.
WiderLns:	Doubles the width of the drawn lines.
NarrwLns:	Halves the width of the drawn lines.
WiderPts:	Doubles the width of the cross of drawn points.
NarrwPts:	Halves the width of the cross of drawn points.
FinrAdapIso:	Doubles the number of adaptive isocurves.
CrsrAdapIso:	Halves the number of adaptive isocurves.
FinerRld:	Doubles number of ruled surfaces in adaptive isocurves.
CrsrRld:	Halves number of ruled surfaces in adaptive isocurves.
RuledSrfApx:	Toggles ruled surface approx. in adaptive isocurves.
AdapIsoDir:	Toggles the row/col direction of adaptive isocurves.
Front:	Selects a front view.
Side:	Selects a side view.
Top:	Selects a top view.
Isometry:	Selects an isometric view.
Clear:	Clears the viewing area.

Obviously not all state options are valid for all drivers. The IRIT server defines in iritinit.irt several user-defined functions that exercise some of the above state commands, such as VIEWSTATE and VIEWSAVE.

In addition to state modification via communication with the IRIT server, modes can be interactively modified on most of the display devices using a pop-up menu that is activated using the right button in the transformation window}. This pop up menu is somewhat different in different drivers, but its entries closely follow the entries of the above state command table.

Animation Mode

All the display drivers are now able to animate objects with animation curves' attributes on them. For more on the way animation curves can be created see the Animation Section of this manual

Once a scene with animation curves' attributes is being loaded into a display device, one can enter "animation" mode using the "Animation" button available in all display devices. The user is then prompt (either graphically or in a textual based interface) for the starting time, termination time and step size of the animation. The parameter space of the animation curve is serving as the time domain. The default staring and terminating times are set as the minimal and maximal parametric domain values of all animation curves. An object at time t below the minimal parametric value will be placed at the starting value of the animation curve. Similarly, an object at time t above the maximal parametric value will be placed at the termination value of the animation curve. The user can also set a bouncing back and forth mode, the number of repetitions, and if desired, request the saving of all the different scenes in the animation as seperate files so a high quality animation can be created.

Specific Comments

The x11drvs supports the following X Defaults (searched at ~/.Xdefaults):

#ifndef COLOR irit*MaxColors: 1 irit*Trans*BackGround: Black irit*Trans*BorderColor: White irit*Trans*TextColor: White irit*Trans*SubWin*BackGround: Black irit*Trans*SubWin*BorderColor: White irit*Trans*CursorColor: White irit*View*BackGround: Black irit*View*BorderColor: White irit*View*CursorColor: White #else irit*MaxColors: 15 irit*Trans*BackGround: NavyBlue irit*Trans*BorderColor: Red irit*Trans*TextColor: Yellow irit*Trans*SubWin*BackGround: DarkGreen irit*Trans*SubWin*BorderColor: Magenta irit*Trans*CursorColor: Green irit*View*BackGround: NavyBlue irit*View*BorderColor: Red irit*View*CursorColor: Red #endif irit*Trans*BorderWidth: 3 irit*Trans*Geometry: =150x500+510+0 irit*View*BorderWidth: 3 irit*View*Geometry: =500x500+0+0

The Motif-based display drivers contain three types of gadgets which can be operated in the following manner. Scales: can be dragged or clicked outside button for single (mouse's middle button) or continuous (mouse's left button) action. Pushbuttons: activated by clicking the mouse's left button. The control panel: allowes rotation, translation of the objects in three axes, determine perspective ratio, viewing object from top, side, front or isometrically, determining scale factor and clipping settings, and operate the matrix stack.

The environment window toggles between screen or object transformation, depth cue on or off, orthographic or perspective projection, wireframe or solid display, single or double buffering, showing or hiding normals, including or excluding the surface;s mesh and curve;s control polygon, surface drawing using isolines or polygons, and four or two polygons per flat patch. Same display devices allow for the inclusion or exclusion of internal edges, and enable or disable of antialiased lines. Scales in the X11/Motif based devices set normals length, lines width, control sensitivity, the number of islolines and samples, etc.

The locations of windows as set via [-g] and [-G] and/or via the configuration file overwrites in x11drvs the Geometry X11 defaults. To use the Geometry X11 default use '-G " "' and '-g " "' or set the string to empty size in the configuration file.

In os2drvs, only -G is used to specify the dimensions of the parent window that holds both the viewing and the transformation window.

In os2drvs, the following key strokes are available as short cuts:

Key	Function
^x	Quit
^s	Save
^f	Front View
^d	Side View
^t	Top View
^i	Isometric VIew
^p	Perspetive/Orthographic
^n	View Internal Edges
^v	View Vertices' Normals
^g	View Polygons' Normals
^b	Backface Culling
^c	Depth Cue
^m	View Control Mesh/Poly

Examples

xglmdrvs -z

prints all the options and their current values.

xglmdrvs -B -i -l 3 solid1.dat

displays the model of solid1.dat using back face culling ('-B'), with internal edges ('-i'), and line width of 3.

xglmdrvs -r -A flat wiggle.dat

displays the model of wiggle.dat shaded ('-r') using flat shading ('-A').

xglmdrvs -I 40 -u -b 255 255 255 wiggle.dat

displays the model of wiggle.dat using isolines' density of 40 ('-I'), using unit matrix to begin with ('-u'), and a white background ('-b').

xglmdrvs -X 0,2,0.1,sx -r anim.dat

executes the animation in anim.dat, from time 0 to time 2 in steps of 0.1. The animation is saved in one frame per file (flag 's' in '-X') and the display device exists once the animation has terminated (flag 'x' in '-X')). The animation will be shaded ('-r').

Utilities - General Usage

The IRIT solid modeler is accompanied by quite a few utilities. They can be subdivided into two major groups. The first includes auxiliary tools such as illustrt and poly3d-h. The second includes filters such as irit2ray and irit2ps.

All these tools operate on input files, and most of the time produce data files. In all utilities that read files, the dash ('-') can be used to read stdin.

Example:

poly3d-h solid1.dat | irit2ps - > solid1.ps

All the utilities have command line options. If an option is set by a '-x' then '-x-' resets the option. The command line options overwrite the settings in config files, and the reset option is useful for cases where the option is set by default, in the configuration file.

All utilities can read a sequence of data files. However, the last transformation matrices found (VIEW_MAT and PRSP_MAT) are actually used.

Example:

poly3d-h solid1.dat | x11drvs solid1.dat - solid1.mat

x11drvs will display the original solid1.dat file with its hidden version, as computed by poly3d-h, all with the solid1.mat, ignoring all other matrices in the data stream.

Under unix, compressed files with a postfix ".Z" will be automatically uncompressed on read and write. The following is legal under unix,

poly3d-h solid1.dat.Z | x11drvs solid1.dat.Z - solid1.mat

where solid1.dat.Z was saved from within IRIT using the command

save( "solid1.dat.Z", solid1 );

or similar. The unix system's "compress" and "zcat" are used for the purpose of (un)compressing the data via pipes. See also SAVE and LOAD.

Poly3d-h - Hidden Line Removing Program

Introduction

poly3d-h is a program to remove hidden lines from a given polygonal model. Freeform objects are preprocessed into polygons with controlled fineness.

The program performs 4 passes over the input:

1. Preprocesses and maps all polygons in a scene, and sorts them.

2. Generates edges out of the polygonal model and sorts them (preprocessing for the scan line algorithm) into buckets.

3. Intersects edges, and splits edges with non-homogeneous visibility (the scan line algorithm).

4. Applies a visibility test of each edge.

This program can handle CONVEX polygons only. From IRIT one can ensure that a model consists of convex polygons only using the CONVEX command:

CnvxObj = convex( Obj );

just before saving it into a file. Surfaces are always decomposed into triangles.

poly3d-h output is in the form of polylines. It is a regular IRIT data file that can be viewed using any of the display devices, for example.

Command Line Options

poly3d-h [-b] [-m] [-i] [-e #Edges] [-H] [-4] [-W Width] [-F PolyOpti FineNess] [-q] [-o OutName] [-t AnimTime] [-c] [-z] DFiles > OutFile

-b: BackFacing - if an object is closed (such as most models created by IRIT), back facing polygons can be deleted, therefore speeding up the process by at least a factor of two.

-m: More - provides some more information on the data file(s) parsed.

-i: Internal edges (created by IRIT) - default is not to display them, and this option will force displaying them as well.

-e n: Number of edges to use from each given polygon (default all). Handy as '-e 1 -4' for freeform data.

-H: Dumps both visible lines and hidden lines as separated objects. Hidden lines will be dumped using a different (dimmer) color and (a narrower) line width.

-4: Forces four polygons per almost flat region in the surface to polygon conversion. Otherwise two polygons only.

-W Width: Selects a default width for visible lines in inches.

-F PolyOpti FineNess: Optimality of polygonal approximation of surfaces. See the variable POLY_APPROX_OPT for the meaning of FineNess. See also -4.

-q: Quiet mode. No printing aside from fatal errors. Disables -m.

-o OutName: Name of output file. Default is stdout.

-t AnimTime: If the data contains animation curves, evaluate and process the scene at time AnimTime.

-z: Prints version number and current defaults.

-c: Clips data to screen (default). If disabled ('-c-'), data outside the view screen ([-1, 1] in x and y) are also processed.

Some of the options may be turned on in poly3d-h.cfg. They can be then turned off in the command line as '-?-'.

Configuration

The program can be configured using a configuration file named poly3d-h.cfg. This is a plain ASCII file you can edit directly and set the parameters according to the comments there. 'poly3d-h -z' will display the current configuration as read from the configuration file.

The configuration file is searched in the directory specified by the IRIT_PATH environment variable. For example, 'setenv IRIT_PATH /u/gershon/irit/bin/'. If the IRIT_PATH variable is not set, the current directory is searched.

Usage

As this program is not interactive, usage is quite simple, and the only control available is using the command line options.

If a certain surface should be polygonized into a finer/caurser set of polygons than the rest of the scene, one can set a "resolution" attribute which specifies the relative FineNess resolution of this specific surface. Further, "u_resolution" and "v_resolution" might be similarly used to set relative resolution for the u or v direction only.

Poly3d-r - A Simple Data Rendering Program

Retired. Sources can be found in the contrib directory, but this program is no longer supported. See irender program instead.

Illustrt - Simple line illustration filter

Introduction

illustrt is a filter that processes IRIT data files and dumps out modified IRIT data files. illustrt can be used to make simple nice illustrations of data. The features of illustrt include depth sorting, hidden line clipping at intersection points, and vertex enhancements. illustrt is designed to closely interact with irit2ps, although it is not neceessary to use irit2ps on illustrt output.

Command Line Options

illustrt [-I #IsoLines] [-f PolyOpti SampTol] [-s] [-M] [-P] [-p] [-O] [-l MaxLnLen] [-a] [-t TrimInter] [-o OutName] [-Z InterSameZ] [-m] [-T AnimTime] [-z] DFiles

-I #IsoLines: Specifies number of isolines per surface, per direction.

-f PolyOpti SampTol: Controls the method used to approximate curves into polylines. If PolyOpti == 0, equally spaced intervals are used. For PolyOpti == 1, an adaptive subdivision that optimizes the samples is employed. For PolyOpti == 2, SampTol (real number) specifies the maximal allowed deviation tolerance of the piecewise linear approximation from the original curve. Default is 0 64 (uniform sampling with 64 samples).

-s: sorts the data in Z depth order that emulates hidden line removal once the data are drawn.

-M: Dumps the control mesh/polygon as well.

-P: Dumps the curve/surface as isocurves.

-p: Dumps vertices of polygons/lines as points.

-O: Handle polygonal objects as possibly open. This will generate two identical edges for an edge shared by two adjacent polygons. Can be useful for open or isilated polygons.

-l MaxLnLen: breaks long lines into shorter ones with maximal length of MaxLnLen. This option is necessary to achieve good depth depending line width in the '-d' option of irit2ps.

-a: takes into account the angle between the two (poly)lines that intersect when computing how much to trim. See also -t.

-t TrimInter: Each time two (poly)line segments intersect in the projection plane, the (poly)line that is farther away from the viewer is clipped TrimInter amount from both sides. See also -a.

-o OutName: Name of output file. Default is stdout.

-Z InterSameZ: The maximal Z depth difference of intersection curves to be be considered invalid.

-m: More talkative mode. Prints processing information.

-T AnimTime: If the data contains animation curves, evaluate and process the scene at time AnimTime.

-z: Prints version number and current defaults.

Usage

illustrt is a simple line illustration tool. It process geometry such as polylines and surfaces and dumps geometry with attributes that will make nice line illustrations. illustrt is geared mainly toward its use with irit2ps to create postscript illustrations. Here is a simple example:

illustrt -s -l 0.1 solid1.dat | irit2ps -W 0.05 -d 0.2 0.6 -u - > solid.ps

make sure all segments piped into irit2ps are shorter than 0.1 and sort them in order to make sure hidden surface removal is correctly applied. Irit2ps is invoked with depth cueing activated, and a default width of 0.05.

illustrt dumps out regular IRIT data files, so output can be handled like any other data set. illustrt does the following processing to the input data set:

Converts surfaces to isocurves ('-I' flag) and isocurves and curves to polylines ('-S' flag), and converts polygons to polylines. Polygonal objects are considered closed and even though each edge is shared by two polygons, only a single one is generated.

Finds the intersection location in the projection plane of all segments in the input data set and trims away the far segment at both sides of the intersection point by an amount controlled by the '-t' and '-a' flags.

Breaks polylines and long lines into short segments, as specified via the '-l' flag, so that width depth cueing can be applied more accurately (see irit2ps's '-d' flag) as well as the Z sorting.

Generates vertices of polygons in the input data set as points in output data controlled via the '-p' flag. set.

Applies a Z sort to the output data, if '-s', so drawing in order of the data will produce a properly hidden surface removal drawing.

Here is a more complex example. Make sure tubular is properly set via "attrib(solid1, "tubular", 0.7);" and invoke:

illustrt -s -p -l 0.1 -t 0.05 solid1.dat | irit2ps -W 0.05 -d 0.2 0.6 -p h 0.05 -u - > solid.ps

makes sure all segments piped into irit2ps are shorter than 0.1, generates points for the vertices, sorts the data in order to make sure hidden surface removal is correctly applied, and trims the far edge by 0.05 at an intersection point. Irit2ps is invoked with depth cueing activated and a default width of 0.05, points are drawn as hollowed circles of default size 0.05, and lines are drawn tubular.

Objects in the input stream that have an attribute by the name of "IllustrtNoProcess" are passed to the output unmodified. Objects in the input stream that have an attribute by the name of "SpeedWave" will have a linear segments added that emulate fast motion with the following attributes,

"Randomness,DirX,DirY,DirZ,Len,Dist,LenRandom,DistRandom,Width".

Objects in the input stream that have an attribute by the name of "HeatWave" will have a spiral curves added that emulate a heat wave in the +Z axis with the following attributes,

"Randomness,Len,Dist,LenRandom,DistRandom,Width".

Examples:

attrib(Axis, "IllustrtNoProcess", ""); attrib(Obj, "SpeedWave", "0.0005,1,0,0,5,3,3,2,0.05"); attrib(Obj, "HeatWave", "0.015,0.1,0.03,0.06,0.03,0.002");

Irender - Simple Scan Line Renderer

Introduction

irender is a program to render IRIT scenes into images. It is a software based Z buffer that is able to create images in few formats. Several of its features includes parametric and volumetric texture mapping, shadow computations, transparency and antialiasing.

Freeform objects are preprocessed into polygons with controlled fineness.

Command Line Options

irender [-z] [-v] [-s XSize YSize] [-a Ambient] [-b R G B] [-B] [-F PolyOpti FineNess] [-f PolyOpti SampTol] [-M Flat/Gouraud/Phong] [-P WMin [WMax]] [-S] [-T] [-t AnimTime] [-A FilterName] [-Z] [-n] [-i rle/ppm] files

-z: Prints version number and current defaults.

-v: Verbose mode. Prints informative messages as it progresses.

-s XSize YSize: Sets the size of the output image, in pixels. Default to 512x512.

-a Ambient: Sets the ambient lighting fraction. Between zero (no ambient lighting) and one. Default to 0.2.

-b R G B: Sets the background color. Each of thre R,G,B colors is an integer value between zero and 255. Default to black.

-B : Apply back face culling. Somewhat faster, but only correct for closed objects. Default is no back face culling.

-F PolyOpti FineNess: Optimality of polygonal approximation of surfaces. See the variable POLY_APPROX_OPT for the meaning of FineNess. Default is 0 and 20.0 (no optimal sampling with fineness of 20.0 (real number)).

-f PolyOpti SampTol: Controls the method used to approximate curves into polylines. If PolyOpti == 0, equally spaced intervals are used. For PolyOpti == 1, an adaptive subdivision that optimizes the samples is employed. For PolyOpti == 2, SampTol (real number) specifies the maximal allowed deviation tolerance of the piecewise linear approximation from the original curve. Default is 0 64 (uniform sampling with 64 samples).

-M Flat/Gouraud/Phong: Selects the shader to be used. Default to Phong if has normals of vertices, Flat if no normals are found.

-P WMin [WMax]: Width of rendered polyline, in pixels. If only WMin is specified, all polylines are set to have WMin width. Otherwise, if WMax is prescribed as well, polylines' width is set to be proportional to their depth with WMax is the width of closest polyline and WMin the farest polyline.

-S: Enable shadow computation. No shadows will be rendered without -S.

-T: Enable transparency computation. No transparent object will be processed without -T.

-t AnimTime: If the data contains animation curves, evaluate and process the scene at time AnimTime.

-A FilterName: Selects an antialiasing filter. FilterName can be one of 'none', 'box', 'triangle', 'quadratic', 'cubic', 'catrom', 'mitchell', 'gaussian', 'sinc, and 'bessel'. Default is 'none'.

-Z: Output will be in the form of Z depth instead of a color image. Output will be 32 bits depth instead of RGBA.

-n: Reverses the normals of vertices and planes, globally.

-i rle/ppm: Selects output image type. Currently the Utah Raster Toolkit's (URT) rle format is being supported as well as the PPM format.

Some of the options may be turned on in irender.cfg. They can be then turned off in the command line as '-?-'.

Configuration

The program can be configured using a configuration file named irender.cfg. This is a plain ASCII file you can edit directly and set the parameters according to the comments there. 'irender -z' will display the current configuration as read from the configuration file.

The configuration file is searched in the directory specified by the IRIT_PATH environment variable. For example, 'setenv IRIT_PATH /u/gershon/irit/bin/'. If the IRIT_PATH variable is not set, the current directory is searched.

Usage

As this program is not interactive, usage is quite simple, and the only control available is using the command line options.

Advanced Usage

One can specify several attributes that affect the way the scene is rendered. The attributes can be generated within IRIT. See also the ATTRIB IRIT command.

Surface color is controlled in two levels. If the object has an RGB attribute, it is used. Otherwise, a color as set via the IRIT COLOR command is used.

If a certain surface should be finer/caurser than the rest of the scene, one can set a "resolution" attribute which specifies the relative FineNess resolution of this specific surface. Further, "u_resolution" and "v_resolution" might be similarly used to set relative resolution for the u or v direction only.

Example:

attrib( Ball, "rgb", "255,0,0" ); color( Sphere, white );

The cosine exponent of the phong shader can be set for a specific object via the SRF_COSINE attribute.

Example:

attrib( Ball, "srf_cosine", 16 );

An object can be drawn transparent instead of opaque, if it has a "transp" attribute. A transparent value of one denotes a completely transparent object, while a value of zero means a completely opaque object. Transparent object will be rendered as such if and only if the '-T' command line option is set.

Example:

attrib( final, "transp", 0.5 );

Several types of texture mapping are supported. Parametric texture may be attached to a parametric surface where the prescribed image is mapped onto the rectangular parametric domain of the surface.

Example:

attrib( Srf1, "ptexture", "checker.ppm" );

The program will automatically detected a ppm file from an rle according to the file's name.

A second type of texture mapping can be applied to all geometric objects. Herein, a procedural texture mapping is employed. Current supported textures are:

wood	Wood style
marble	Marble style
contour	Parallel plane contouring
ncontour	Constant normal angle to major axis

A second parameter that must be provided for procedural textures is the scaling factor of the texture, which can be either one parameter of uniform scaling or a vector of three coefficients for scaling in x, y, and z. The two parameters are separated by a comma. For contour style, the scale denotes the spacing of the contouring planes in X, Y and Z. For ncontour style, the scale also denotes the spacing of the adjacent constant normal contours. Related attributes are "texture_color" and "texture_width" that supports the color and the width of the textured strokes.

Example:

attrib( Obj1, "texture", "marble, 2" ); attrib( Obj2, "texture", "contour, 1 0.5 2.5" ); attrib( Obj2, "texture_width", "0.05" ); attrib( Obj2, "texture_color", "255,255,255" );

which sets Obj1 to have a marble procedural texture with a uniform scaling factor of 2 and a contouring texture for Obj2 with scaling factors of (1, 0.5, 2.5) in x, y, and z, in white color and width 0.05.

In addition, a scalar surface spanning the same parameteric domain as an original surface may be used as texture mapping function. Herein, the scalar function texture is evaluated at each UV parameter value and is mapped through a color scale to yield the output color. This type of texture is useful for stress maps or analysis maps on top of freeform surfaces. Several related attributes are supported: "stexture_scale" which prescribes the color scale image (only its first column is employed), and "stexture_bound" that sets the domain that will be clipped to the min max values. Funally, "stexture_func" can hold the functions "sqrt" or "abs" to be applied to the evaluated surface value.

Example:

attrib( Srf, "stexture", scrvtr( Srf, P1, off ) ); attrib( Srf, "stexture_scale", "color_scale.ppm" ); attrib( Srf, "stexture_func", "sqrt" ); attrib( Srf, "stexture_bound", "0.0 100.0" );

where scrvtr computes a scalar field to Srf that represents the sum of the squares of the principle curvatures. The evaluated scalar texture surface's value is piped through a sqrt function. The first column of the image of color_scale.ppm is used to set the coloring scale for curvature bounds values between 0.0 and 100.0.

Both "stexture_scale" and "stexture_bound" are optional. The default color scale maps the min/max values from blue to red through green. The default scalar surface texture bound is computed as the extreme values of the "stexture" surface.

While the program has a default for lighting which is two light sources at opposite directions at (1, 1, 1) and (-1, -1, -1), one can overwrite this default. A POINT_TYPE object with LIGHT_SOURCE attribute denotes a light source. If irender detects one or more light sources in the input stream, the default light sources are not created. Two types of light sources may be prescribed, a parallel at infinity or a point at finite distance light source, distinguished by a TYPE attribute of either POINT_POS or POINT_INFTY. A point light source can be colored, when an RGB attribute will set its color. A point light source will cast shadows if and only if it has SHADOW attribute (one needs to apply the '-S' command line option as well for rendering shadows). Finally, one can construct two mirrored light sources at opposite directions if TWOLIGHT attribute is added to the light source object.

Example:

Light1 = point( 0, 0, 10 ); attrib( Light1, "light_source", on ); attrib( Light1, "shadow", on ); attrib( Light1, "rgb", "255,0,0" ); attrib( Light1, "type", "point_pos" ); Light2 = point( 1, 1, 1 ); attrib( Light2, "light_source", on ); attrib( Light2, "twolight", on ); attrib( Light2, "type", "point_infty" );

constructs two lights sources with Light1 with red color positioned at (0, 0, 10) and casting shadows, while Light2 will create two mirrored white parallel lights sources in the direction of (1, 1, 1) and (-1, -1, -1), as is irender's default.

Command Line Options

irit2nff [-l] [-4] [-c] [-F PolyOpti FineNess] [-o OutName] [-T] [-t AnimTime] [-g] [-z] DFiles

-l: Linear - forces linear (degree two) surfaces to be approximated by a single polygon along their linear direction. Although, most of the time, linear direction can be exactly represented using a single polygon, even a bilinear surface can have a free-form shape (saddle-like) that is not representable using a single polygon. Note that although this option will better emulate the surface shape, it will create unnecessary polygons in cases where one is enough.

-4: Four - Generates four polygons per flat patch. Default is 2.

-c: Output files should be filtered by cpp. When set, the usually huge geometry file is separated from the main nff file that contains the surface properties and view parameters. By default all data, including the geometry, are saved into a single file with type extension '.nff'. Use of '-c' will pull out all the geometry into a file with the same name but a '.geom' extension, which will be included using the '#include' command. The '.nff' file should, in that case, be preprocessed using cpp before being piped into the nff renderer.

-F PolyOpti FineNess: Optimality of polygonal approximation of surfaces. See the variable POLY_APPROX_OPT for the meaning of FineNess. See also -4.

-o OutName: Name of output file. By default the name of the first data file from the DFiles list is used. See below on the output files.

-T: Talkative mode. Prints processing information.

-t AnimTime: If the data contains animation curves, evaluate and process the scene at time AnimTime.

-g: Generates the geometry file only. See below.

-z: Prints version number and current defaults.

Usage

Irit2Nff converts freeform surfaces into polygons in a format that can be used by an NFF renderer. Usually, one file is created with '.nff' type extension. Since the number of polygons can be extremely large, a '-c' option is provided, which separates the geometry from the surface properties and view specification, but requires preprocessing by cpp. The geometry is isolated in a file with extension '.geom' and included (via '#include') in the main '.nff' file. The latter holds the surface properties for all the geometry as well as the viewing specification. This allows for the changing of shading or the viewing properties while editing small ('.nff') files.

If '-g' is specified, only the '.geom' file is created, preserving the current '.nff' file. The '-g' flag can be specified only with '-c'.

In practice, it may be useful to create a low resolution approximation of the model, change viewing/shading parameters in the '.nff' file until a good view and/or surface quality is found, and then run Irit2Nff once more to create a high resolution approximation of the geometry using '-g'.

Example:

irit2nff -c -l -F 0 8 b58.dat

creates b58.nff and b58.geom with low resolution (FineNess of 5).

Once done with parameter setting, a fine approximation of the model can be created with:

irit2nff -c -l -g -F 0 64 b58.dat

which will only recreate b58.geom (because of the -g option).

One can overwrite the viewing matrix by appending a new matrix in the end of the command line, created by a display device:

xgldrvs b58.dat irit2nff -l -F 0 32 b58.dat irit.mat

where irit.mat is the viewing matrix created by xgldrvs.

Advanced Usage

Irit2Scn - IRIT To SCENE (RTrace) filter

SCENE is the format used by the RTrace ray tracer. This filter was donated by Antonio Costa (acc@asterix.inescn.pt), the author of RTrace.

Command Line Options

irit2scn [-l] [-4] [-F PolyOpti FineNess] [-o OutName] [-g] [-T] [-t AnimTime] [-z] DFiles

-l: Linear - forces linear (degree two) surfaces to be approximated as a single polygon along their linear direction. Although most of the time, linear direction can be exactly represented using a single polygon, even a bilinear surface can have a free-form shape (saddle-like) that is not representable using a single polygon. Note that although this option will better emulate the surface shape, it will create unnecessary polygons in cases where one is enough.

-4: Four - Generates four polygons per flat patch.

-F PolyOpti FineNess: Optimality of polygonal approximation of surfaces. See the variable POLY_APPROX_OPT for the meaning of FineNess. See also -4.

-o OutName: Name of output file. By default the name of the first data file from DFiles list is used. See below on the output files.

-g: Generates the geometry file only. See below.

-T: Talkative mode. Prints processing information.

-t AnimTime: If the data contains animation curves, evaluate and process the scene at time AnimTime.

-z: Prints version number and current defaults.

Usage

Irit2Scn converts freeform surfaces and polygons into polygons in a format that can be used by RTrace. Two files are created, one with a '.geom' extension and one with '.scn'. Since the number of polygons can be extremely large, the geometry is isolated in the '.geom' file and is included (via '#include') in the main '.scn' file. The latter holds the surface properties for all the geometry as well as viewing and RTrace specific commands. This allows for the changing of the shading or the viewing properties while editing small ('.scn') files.

If '-g' is specified, only the '.geom' file is created, preserving the current '.scn' file.

In practice, it may be useful to create a low resolution approximation of the model, change the viewing/shading parameters in the '.scn' file until a good view and/or surface quality is found, and then run Irit2Scn once more to create a high resolution approximation of the geometry using '-g'.

Example:

irit2scn -l -F 0 8 b58.dat

creates b58.scn and b58.geom with low resolution (FineNess of 5).

One can ray trace this scene after converting the scn file to a sff file, using scn2sff provided with the RTrace package.

Once done with the parameter setting of RTrace, a fine approximation of the model can be created with:

irit2scn -l -g -F 0 64 b58.dat

which will only recreate b58.geom (because of the -g option).

One can overwrite the viewing matrix by appending a new matrix in the end of the command line, created by the display devices:

wntdrvs b58.dat irit2scn -l -F 0 8 b58.dat irit.mat

where irit.mat is the viewing matrix created by wntdrvs. The output name, by default, is the last input file name, so you might want to provide an explicit name with the -o flag.

Advanced Usage

One can specify surface qualities for individual surfaces of a model. Several such attributes are supported by Irit2Scn and can be set within IRIT. See also the ATTRIB IRIT command.

If a certain surface should be finer/caurser than the rest of the scene, one can set a "resolution" attribute which specifies the relative FineNess resolution of this specific surface. Further, "u_resolution" and "v_resolution" might be similarly used to set relative resolution for the u or v direction only.

Example:

attrib( srf1, "resolution", 2 );

will force srf1 to have twice the default resolution, as set via the '-f' flag.

Almost flat patches are converted to polygons. The patch can be converted into two polygons (by subdividing along one of its diagonals) or into four by introducing a new point at the patch center. This behavior is controlled by the '-4' flag, but can be overwritten for individual surfaces bu setting "twoperflat" or "fourperflat".

RTrace specific properties are controlled via the following attributes: "SCNrefraction", "SCNtexture", "SCNsurface. Refer to the RTrace manual for their meaning.

Example:

attrib( srf1, "SCNrefraction", 0.3 );

Surface color is controlled in two levels. If the object has an RGB attribute, it is used. Otherwise a color as set via IRIT COLOR command is used if set.

Example:

attrib( tankBody, "rgb", "244,164,96" );

Irit2Xfg - IRIT To XFIG filter

Command Line Options

irit2xfg [-s Size] [-t XTrans YTrans] [-I #UIso[:#VIso[:#WIso]]] [-f PolyOpti SampTol] [-F PolyOpti FineNess] [-M] [-G] [-T] [-a AnimTime] [-i] [-o OutName] [-z] DFiles

-s Size: Size in inches of the page. Default is 7 inches.

-t XTrans YTrans: X and Y translation. of the image. Default is (0, 0).

-I #UIso[:#VIso]: Specifies the number of isolines per surface, per direction. If #VIso is not specified, #UIso is used for #VIso as well.

-f PolyOpti SampTol: Controls the method used to approximate curves into polylines. If PolyOpti == 0, equally spaced intervals are used. For PolyOpti == 1, an adaptive subdivision that optimizes the samples is employed. For PolyOpti == 2, SampTol (real number) specifies the maximal allowed deviation tolerance of the piecewise linear approximation from the original curve. Default is 0 64 (uniform sampling with 64 samples).

-F PolygonOpti FineNess: Optimality of polygonal approximation of surfaces. See the variable POLY_APPROX_OPT for the meaning of FineNess. See also -4. This enforces the dump of freefrom geometry as polygons.

-M: Dumps the control mesh/polygon as well.

-G: Dumps the freeform geometry.

-T: Talkative mode. Prints processing information.

-a AnimTime: If the data contains animation curves, evaluate and process the scene at time AnimTime.

-i: Internal edges (created by IRIT) - default is not to display them, and this option will force displaying them as well.

-o OutName: Name of output file. By default the name of the first data file from DFiles list is used. See below on the output files.

-z: Prints version number and current defaults.

Usage

Irit2Xfg converts freeform surfaces and polygons into polylines in a format that can be used by XFIG.

Example:

irit2Xfg -T -f 0 16 saddle.dat > saddle.xfg

However, one can overwrite the viewing matrix by appending a new matrix in the end of the command line, created by the display devices:

x11drvs b58.dat irit2Xfg -T -f 0 16 b58.dat irit.mat > saddle.xfg

where irit.mat is the viewing matrix created by x11drvs.

Obj2irit - Wavefront OBJ format To IRIT data files

Converts Waverfront's OBJ data files into IRIT data files.

Command Line Options

obj2irit [-m] [-o OutName] [-z] OBJFile

-m: Provides some more information on the data file(s) parsed.

-o OutName: Name of output file. By default the output goes to stdout.

-z: Prints version number and current defaults.

Usage

obj2irit converts Wavefront's OBJ data files into IRIT data files. The current version provides only partial support for the direct conversion of freeform surfaces, mainly due to luck of examples of freeform surfaces in obj format.

Example:

obj2irit -o file.dat file.obj

Data File Format

This section describes the data file format used to exchange data between IRIT and its accompanying tools.

[OBJECT {ATTRS} OBJNAME [NUMBER n] | [VECTOR x y z] | [CTLPT POINT_TYPE {w} x y {z}] | [STRING "a string"] | [MATRIX m00 ... m03 m10 ... m13 m20 ... m23 m30 ... m33] ;A polyline should be drawn from first point to last. Nothing is drawn ;from last to first (in a closed polyline, last point is equal to first). | [POLYLINE {ATTRS} #PTS ;#PTS = number of points. [{ATTRS} x y z] [{ATTRS} x y z] . . . [{ATTRS} x y z] ] ;Defines a closed planar region. Last point is NOT equal to first, ;and a line from last point to first should be drawn when the boundary ;of the polygon is drawn. | [POLYGON {ATTRS} #PTS [{ATTRS} x y z] [{ATTRS} x y z] . . . [{ATTRS} x y z] ] ;Defines a "cloud" of points. | [POINTLIST {ATTRS} #PTS [{ATTRS} x y z] [{ATTRS} x y z] . . . [{ATTRS} x y z] ] ;Defines an instance - a geometry reference (by name, SRF13 below) ;and a transformation matrix to apply to this geoemtry | [INSTANCE SRF13 m00 ... m03 m10 ... m13 m20 ... m23 m30 ... m33] ;Defines a Bezier curve with #PTS control points. If the curve is ;rational, the rational component is introduced first. | [CURVE BEZIER {ATTRS} #PTS POINT_TYPE [{ATTRS} {w} x y z ...] [{ATTRS} {w} x y z ...] . . . [{ATTRS} {w} x y z ...] ] ;Defines a Bezier surface with #UPTS * #VPTS control points. If the ;surface is rational, the rational component is introduced first. ;Points are printed row after row (#UPTS per row), #VPTS rows. | [SURFACE BEZIER {ATTRS} #UPTS #VPTS POINT_TYPE [{ATTRS} {w} x y z ...] [{ATTRS} {w} x y z ...] . . . [{ATTRS} {w} x y z ...] ] ;Defines a Bezier triangular surface with (#PTS + 1) * #PTS / 2 control ;points, of order ORDER. If the surface is rational, the rational ;component is introduced first. Note #PTS holds number of points along ;an edge and is exactly equal to ORDER. Points are printed sequentially. | [TRISRF BEZIER {ATTRS} #PTS POINT_TYPE [{ATTRS} {w} x y z ...] [{ATTRS} {w} x y z ...] . . . [{ATTRS} {w} x y z ...] ] ;Defines a Bezier trivariate with #UPTS * #VPTS * #WPTS control ;points. If the trivariate is rational, the rational component is ;introduced first. Points are printed row after row (#UPTS per row), ;#VPTS rows, #WPTS layers (depth). | [TRIVAR BEZIER {ATTRS} #UPTS #VPTS #WPTS POINT_TYPE [{ATTRS} {w} x y z ...] [{ATTRS} {w} x y z ...] . . . [{ATTRS} {w} x y z ...] ] ;Defines a Bspline curve of order ORDER with #PTS control points. If the ;curve is rational, the rational component is introduced first. ;Note length of knot vector is equal to #PTS + ORDER. ;If curve is periodic KVP prefix the knot vector that has length of ;'Length + Order + Order - 1'. | [CURVE BSPLINE {ATTRS} #PTS ORDER POINT_TYPE [KV{P} {ATTRS} kv0 kv1 kv2 ...] ;Knot vector [{ATTRS} {w} x y z ...] [{ATTRS} {w} x y z ...] . . . [{ATTRS} {w} x y z ...] ] ;Defines a Bspline surface with #UPTS * #VPTS control points, of order ;UORDER by VORDER. If the surface is rational, the rational component ;is introduced first. ;Points are printed row after row (#UPTS per row), #VPTS rows. ;If surface is periodic in some direction KVP prefix the knot vector ;that has length of 'Length + Order + Order - 1'. | [SURFACE BSPLINE {ATTRS} #UPTS #VPTS UORDER VORDER POINT_TYPE [KV{P} {ATTRS} kv0 kv1 kv2 ...] ;U Knot vector [KV{P} {ATTRS} kv0 kv1 kv2 ...] ;V Knot vector [{ATTRS} {w} x y z ...] [{ATTRS} {w} x y z ...] . . . [{ATTRS} {w} x y z ...] ] ;Defines a Bspline triangular surface with (#PTS + 1) * #PTS / 2 control ;points, of order ORDER. If the surface is rational, the rational ;component is introduced first. ;Points are printed sequentially. | [TRISRF BSPLINE {ATTRS} #PTS ORDER POINT_TYPE [KV {ATTRS} kv0 kv1 kv2 ...] ;Knot vector [{ATTRS} {w} x y z ...] [{ATTRS} {w} x y z ...] . . . [{ATTRS} {w} x y z ...] ] ;Defines a Bspline trivariate with #UPTS * #VPTS * #WPTS control ;points. If the trivariate is rational, the rational component is ;introduced first. Points are printed row after row (#UPTS per row), ;#VPTS rows, #WPTS layers (depth). ;If trivariate is periodic in some direction KVP prefix the knot vector ;that has length of 'Length + Order + Order - 1'. | [TRIVAR BSPLINE {ATTRS} #UPTS #VPTS #WPTS UORDER VORDER WORDER POINT_TYPE [KV{P} {ATTRS} kv0 kv1 kv2 ...] ;U Knot vector [KV{P} {ATTRS} kv0 kv1 kv2 ...] ;V Knot vector [KV{P} {ATTRS} kv0 kv1 kv2 ...] ;W Knot vector [{ATTRS} {w} x y z ...] [{ATTRS} {w} x y z ...] . . . [{ATTRS} {w} x y z ...] ] ;Defines a trimmed surface. Encapsulates a surface (can be either a ;Bspline or a Bezier surface) and prescribes its trimming curves. ;There can be an arbitrary number of trimming curves (either Bezier ; or Bspline). Each trimming curve contains an arbitrary number of ;trimming curve segments, while each trimming curve segment contains ;a parameteric representation optionally followed by a Euclidean ;representation of the trimming curve segment. | [TRIMSRF [SURFACE ... ] [TRIMCRV [TRIMCRVSEG [CURVE ... ] ] . . . [TRIMCRVSEG [CURVE ... ] ] ] . . . [TRIMCRV [TRIMCRVSEG [CURVE ... ] ] . . . [TRIMCRVSEG [CURVE ... ] ] ] ] ] POINT_TYPE -> E1 | E2 | E3 | E4 | E5 | P1 | P2 | P3 | P4 | P5 ATTRS -> [ATTRNAME ATTRVALUE] | [ATTRNAME] | [ATTRNAME ATTRVALUE] ATTRS

Some notes:

* This definition for the text file is designed to minimize the reading time and space. All information can be read without backward or forward referencing.

* An OBJECT must never hold different geometry types or other entities. I.e. CURVEs, SURFACEs, and POLYGONs must all be in different OBJECTs.

* Attributes should be ignored if not needed. The attribute list may have any length and is always terminated by a token that is NOT 'verb+[+'. This simplifies and disambiguates the parsing.

* Comments may appear between 'verb+[+OBJECT ...verb+]+' blocks, or immediately after OBJECT OBJNAME, and only there.

A comment body can be anything not containing the 'verb+[+' or the 'verb+]+' tokens (signals start/end of block). Some of the comments in the above definition are illegal and appear there only of the sake of clarity.

* It is preferred that geometric attributes such as NORNALs will be saved in the geometry structure level (POLYGON, CURVE or vertices) while graphical and others such as COLORs will be saved in the OBJECT level.

* Objects may be contained in other objects to an arbitrary level.

Here is an example that exercises most of the data format:

This is a legal comment in a data file. [OBJECT DEMO [OBJECT REAL_NUM And this is also a legal comment. [NUMBER 4] ] [OBJECT A_VECTOR [VECTOR 1 2 3] ] [OBJECT CTL_POINT [CTLPT E3 1 2 3] ] [OBJECT STR_OBJ [STRING "string"] ] [OBJECT UNIT_MAT [MATRIX 1 0 0 0 0 1 0 0 0 0 1 0 0 0 0 1 ] ] [OBJECT [COLOR 4] POLY1OBJ [POLYGON [PLANE 1 0 0 0.5] 4 [-0.5 0.5 0.5] [-0.5 -0.5 0.5] [-0.5 -0.5 -0.5] [-0.5 0.5 -0.5] ] [POLYGON [PLANE 0 -1 0 0.5] 4 [0.5 0.5 0.5] [-0.5 0.5 0.5] [-0.5 0.5 -0.5] [0.5 0.5 -0.5] ] ] [OBJECT [COLOR 63] ACURVE [CURVE BSPLINE 16 4 E2 [KV 0 0 0 0 1 1 1 2 3 4 5 6 7 8 9 10 11 11 11 11] [0.874 0] [0.899333 0.0253333] [0.924667 0.0506667] [0.95 0.076] [0.95 0.76] [0.304 1.52] [0.304 1.9] [0.494 2.09] [0.722 2.242] [0.722 2.318] [0.38 2.508] [0.418 2.698] [0.57 2.812] [0.57 3.42] [0.19 3.572] [0 3.572] ] ] [OBJECT [COLOR 2] SOMESRF [SURFACE BEZIER 3 3 E3 [0 0 0] [0.05 0.2 0.1] [0.1 0.05 0.2] [0.1 -0.2 0] [0.15 0.05 0.1] [0.2 -0.1 0.2] [0.2 0 0] [0.25 0.2 0.1] [0.3 0.05 0.2] ] ] ]

Bugs and Limitations

Like any program of more than one line, it is far from being perfect. Some limitations, as well as simplifications, are laid out below.

* If the intersection curve of two objects falls exactly on polygon boundaries, for all polygons, the system will scream that the two objects do not intersect at all. Try to move one by EPSILON into the other. I probably should fix this one - it is supposed to be relatively easy.

* Avoid degenerate intersections that result with a point or a line. They will probably cause wrong propagation of the inner and outer part of one object relative to the other. Always extend your object beyond the other object.

* If two objects have no intersection in their boundary, IRIT assumes they are disjoint: a union simply combines them, and the other Boolean operators return a NULL object. One should find a FAST way (3D Jordan theorem) to find the relation between the two (A in B, B in A, A disjoint B) and according to that, make a decision.

* Since the boolean sum implementation constructs ruled surfaces with uniform speed, it might return a somewhat incorrect answer, given non-uniform input curves.

* The parser is out of hand and is difficult to maintain. There are several memory leaks there that one should fix.

* The X11 driver has no menu support (any easy way to have menus using Xlib!?).

* IBM R6000 fails to run the drivers in -s- mode.

* Rayshade complains a lot about degenerate polygons on irit2ray output. To alleviate the problem, change the 'equal' macro in common.h in libcommon of rayshade from EPSILON (1e-5) to 1e-7 or even lower.

* On the motif-based drivers (xmtdrvs etc.) clicking the mouse left and right of the scale's button produces stepped transformations. This step size is constant, and is not proportional to the distance between the mouse's position and the position of the button. The reason for the flaw is incorrect callback information returned from the scale in repeattive mode.

* Binary data files are not documented, nor will they be. They might change in the future and are in fact machine dependend. Hence, one platform might fail to read the other's binary data file.