FishMarket 1.0/FishMarket v1.0.iso/fishies/426-450/disk_446/post/readme

Text File | 1992-05-06 | 17KB | 356 lines

Postscript interpreter - README
===============================

Post V1.4 Copyright Adrian Aylward 1989, 1991

This distribution was released as the archive file "post13.lzh", ??-???-90

Files in this distribution:

README            This file
Post.info         Icon file
post              User interface
post.doc          Documentation file
postband          Band printing driver
postband.doc      Documentation file
postlj            LaserJet printing driver
postlj.doc        Documentation file
post.library      Shared library for Amiga 500/1000/2000
post.library.2620 Shared library for Amiga 2500 etc. (68020/030, 68881/2)
init.ps           Standard postscript startup file
icon.ps           The postscript program that built the icon

Source files (user interface only), for programmers:

post.c            User interface, C main program
postband.c        Band printing driver, C main program
postlj.c          LaserJet printing driver, C main program
postasm.a         User interface, assembler support
postlib.h         Library interface header
postlib.doc       Documentation file
makefile          Make file

You will also need:

arp.library       Arp library, V39+
conman            ConMan, V1.3+

You should find these on "All Good BBS's/disk libraries"

I have tested Post under WorkBench V1.3 only.  If you have 2.0 it should
work, but you may need a more recent version of ConMan.  You may also need
to use the CONDEV option to use ConMan instead of the standard console
handler.

You will need at least a megabyte of memory.  Even with a megabyte you
will have not be able to generate a full page in colour on the screen.
For high density printer output (A4 at 300 dpi.) you will need 2 megabytes!  If memory is very tight, try running from the CLI, before workbench has
been loaded.

Before using Post you will need to copy the library (post.library or
post.library.2620) into your libs: directory.  If you have a 32 bit cpu
(68020/030/040) and a floating point unit (68881/68882 or 68040) you can
use the 2620 version of the library, which is compiled to use hardware
floating point instructions and runs about twice as fast.  Otherwise you
must use the standard library - if you try the 2620 version it will
refuse to open.

Examples
========

The examples and demo programs are now being distributed separately.  Look
out for the file "psdemos1.lzh".  (If you have the prerelease "post02.lzh"
or one of the earlier releases there is no need to download them, as they
have not changed significantly).

The PostScript Language
=======================

If you don't know any PostScript I recomend obtaining a suitable tutorial
book.  Try the "blue book" from Adobe:

    "PostScript Language Tutorial and Cookbook"
    Adobe Systems Incorporated,
    Addison-Wesley 1985, ISBN 0-201-10179-3

If you want to learn PostScript, Post is a valuable learning tool, as you
run it interactively and view the output one the screen.  Without some basic
knowledge you may find it difficult to understand what is happening if you
have problems getting started.

The startup file
================

The standard startup file "init.ps" should be run before most programs.  You
should read it and the comments within it, as you may wish to customise it.
It defines useful operators to mimic a standard PostScript printer.  It also
handles automatic font downloading (see below).

Getting started
===============

Make sure that you have ConMan (V1.3+) and the ARP library installed.

Copy the file post.library into your libs: directory.  If you have a
68020 or 68030 cpu and a 68881 or 6882 floating unit you can use the 2620
version of the library instead - copy it to libs:post.library.

From the CLI you can just type "post init.ps screen", assuming that all the
files are in the current directory.  You should see a new screen with a big
requestor window appear.  Click on "OK" and wait while Post executes the
startup file, then you are ready to go.

If it says that it can't get enough memory (if you have only 1 megabyte)
try selecting "Black and White" on the requestor window, and selecting
the "Close WBench" option.

For a quick check of the graphics, pick File/Interactive on the menu.  Then
type into the narrow window at the top:

    200 200 150 0 360 arc stroke

The result should be a circle appearing in the lower window.  When you have
finished playing, type "quit".  Then "Project/Quit" will return you to the
CLI.

If you don't have any fonts yet all text output will just print as big dots.
(Read the section below on fonts).  Make a "PSFonts:" directory and copy all
your PostScript fonts into it.  (Don't confuse these with Amiga bitmapped
fonts - these will NOT work, nor will .AFM font metric files)  Their names
should be the same as their PostScript names - e.g. "PSFonts:Times-Roman",
"PSFonts:Times-Bold".  The files can either be ASCII text or in IBM
PostScript font format.  If you have copied them from an IBM disk you will
have to change the names; since MSDOS can only handle 8 character names its
stores Times-Roman as "TI______.PFB" and Times-Bold as "TIB_____.PFB".  Then
Post will load the fonts automatically as they are needed.

To use the workbench startup click on the icon.  You will probably want to
set up the tooltypes to execute the startup file "init.ps"; for now you can
enter it in the requestor.

Fonts
=====

As there are no fonts built in to the interpreter, all fonts must be loaded
from the disk, as PostScript source.  The standard startup file redefines
the findfont operator to automatically load fonts from the disk the first
time they are referenced.  To use this, build a directory "PSFonts:"
containing your PostScript fonts.  The file names should be the same as the
PostScript names, without the '/'.  You can add an extension if you wish,
by editing the startup file.  You can also make it search more than one
directory.

The startup file also defines a default font, which will be substituted if
the font your program requests cannot be found.  The initial default font
as set up by the standard startup file is a dummy font that prints big dots,
but at least allows most text operations to proceed without producing errors.
This means that if you don't have the font loading set up properly the only
text output you will get is big dots.  So if that is what you see when you
run Post it probably means that you don't have any fonts, rather than a
problem with Post itself.

Since a font is loaded into the vm just like any other PostScript code, it
will be discarded when the vm is restored.  This means that in a multiple
page document the fonts will likely be reloaded on every page.  To save time,
it is usually much better to preload the fonts you are going to need.  You
can do this using the "load font" menu command, or you can write yourself
a little PostScript program to load your favourite fonts.  (Exercise for
the user: write a driver to extract the font names from a conforming
PostScript program and preload them.)  If you are using a large number of
different fonts at once you may need to increase the default vm size to
contain them.

There are a number of commercial sources for fonts.  Both the Adobe type 1
(encrypted) fonts and type 3 (standard downloadable) fonts are supported.
Since type 1 fonts can contain special "hints" to improve the quality at
small point sizes on low or medium resolution devices they usually give
much higher quality results.  You can get type 1 fonts from Adobe, or a
number of other vendors.

A good source of the standard set of 35 fonts as come on most PostScript
printers is the Adobe Type Manager (13 fonts) and its associated Plus Pack
(22 fonts).  The IBM version contains the fonts in a format which can be read
by Post.  Various utility programs (e.g. CROSSDOS) are available to copy file
from IBM 3.5 inch disks onto Amiga files.  Just copy the .PFB files into your
PSFonts: directory, changing the names to their PostScript names. The Mac
version contains the fonts too, but you need a utility program to change the
file format.  The quality of the Adobe fonts is excellent; the printed
output from Post on my LJIIP rivals that of a LaserWriter.

If you have any of the additional CompuGraphic outline fonts supplied by
Gold Disk for use with their DTP programs you can generate type 3 fonts
from them using their conversion program.  The quality is not as good as
the original Compugraphic fonts but is acceptable for proofing purposes.

There are some freely copyable fonts tht were originally distributed with
the GhostScript project (another freeware PostScript interpreter available
on Unix).  The quality of these is not high, as they appear to have been
generated by tracing bitmap outlines, but they will serve to get you
started, and they are free.

There are also various shareware and PD fonts available.  Many of these were
derived from the Mac or IBM.  They may still be in a foreign file format.
The file should be converted to readable ASCII, with no strange control
characters. Some of the fonts have Mac encoding vectors.  You won't notice
the difference for the standard ASCII characters, as they are the
same, but you may have to convert to the standard encoding to get
more unusual characters like the trademark sign to work.  (Programs
like Adobe illustrator do their own font encoding anyway.)  You
will need to know some PostScript to make this change.  If you
see the sequence "/Helvetica findfont /Encoding get" or
"<dictionary> /MacVec get" try replacing it with "StandardEncoding".

You are recommended to ensure all your fonts have UniqueID's.  This makes
the font caching code more efficient.  Often all you have to do is uncomment
them in the source.  Any number will do for the ID (up to 2**24 - 1), as long
as it is different for each of your fonts.  Make it a multiple of 4, as Adobe
recommend adding 1 or 2 to the ID when modifying the font.

Each font should have a FontName entry.  This isn't actually required by
Adobe, but certain programs such as Adobe Illustrator copy the font directory
when recoding the font.  If the name is missing, you get a dictionary full
error after the name has been added.

WorkBench 2.0
=============

I don't have 2.0 myself, so I am not able to test Post under it.  But I
understand that it works OK.  You need to set up ConMan it you want to
run interactively.  It you mount it as "CNC:" the you can use the CONDEV
option to access it.  Otherwise you will find the console output is going
to the WorkBench screen.

Common problems
===============

Printers
--------

Some people have difficulty in getting Post to work correctly with printer
output.  The first thing to check is that you have your printer preferences
set up properly.  Post uses the printer values to set the size and density
of its page buffer, so if you have silly values set it may pick a silly
value for its buffer size and fail due to running out of memory.  So Run
Preferences, go to the graphic 2 screen, and check the page limits.  If you
set the limits to "Absolute" then width and height should be in inches; if
you choose "Pixels" then the dimensions are in dots.  Turn smoothing and
centering off.  On the graphic 1 screen, set the Aspect to horizontal, the
image to positive, and the shade to black and white or colour according to
your printer.  If the output comes out distorted, to a silly scale, or with
circles turned into ellipses, and takes forether to print, it is probably
because the page size is wrong and the printer device is laboriously scaling
it to fit.

If you are having problems with the printer setup, it is better to run Post
interactively.  Then you can look at the values in the page size gadgets
in the requestor window to check they are all sensible.

Most page printers (lasers, DeskJet) will give the best quality with density
set to the highest value.  For matrix printers try both the highest and the
next highest - as the pins usually overlap at high densities the quality may
not be any better.  If you are very tight on memory then a lower density
will reduce the amount needed for the page buffer.

The Workbench printer drivers print bitmapped graphics dumps within the
text margins, whereas PostScript expects it bitmaps to begin right at the
edges of the paper.  You can line things up properly by setting the page
offset gadgets in the requestor window.  Once you have experimented to
determine the proper values to can put them in a shell script, or set them
in the WorkBench icon, so you don't have to type them every time.

If you have a LaserJet printer or one that emulates it, you should use the
special driver "postlj".  It is faster, supports all the paper sizes in
both posrtrait and landscape orientation, and avoids all the problems with
the standard Workbench drivers interface.  I have tested it on my HP LJIIP,
but it should work on other LaserJet compatibles.

Fonts
-----

Most of the remaining troubles relate to fonts.  I personally recommend
the Adobe fonts: the quality is excellent and you are assured of maximum
compatability.  There are many other sources for fonts.  Even the shareware
fonts normally work.  However being constructed slightly differently from the
Adobe ones, they sometimes fail when programs attempt to modify them to
change the character set encoding or character widths or obtain special
effects.

Other troubles with fonts usually relate to not not being able to find the
font file on the disk.  Read the comments in the init.ps file to see
what directories in you filing system it is searching, also check whether
you font files have extensions.  I prefer to keep my fonts without
extensions, so for example on my machine "Times-Roman" is in the file
"PSFonts:Times-Roman".  But you may want to keep the ".PFB" extension on
your font filesif they are in IBM format - if you have other packages that
like them that way.  You can easily change this by editing init.ps.

N.B. if you have some files with ".AFM" extensions they are NOT PostScript
fonts.  The are font metric files that define the character widths and other
data for use by application programs; they do not define the characters
outlines themselves.  Ther are of no use to Post.

Use with other packages
-----------------------

I have used Post extensively with PPage (V1.3).  Early versions of PPage
had a bug in the PostScript prologue that caused a limitcheck error when
doping patterened fills, but this has been fixed for a while now.  I find
that printing using postlj is about five times faster than printing direct
from PPage.  It solves the page offset problem.  Also, it gives access to
the Adobe type 1 font range, which is considerably more extensive than the
hinted versions of the Compugraphic fonts that are presently available.

I have tested a PageStream file.  There was a problem with the prologue
not being compatable with downloaded Adobe fonts.  If you load all your
fonts beforehand, using the "Load font" menu selection, it should work OK.
Hopefully this will eb fixed soon.

I have printed Excellence! files successfully.

AmigaTeX with the PostScript driver should work OK.  I have seen a TeX
example (from another system) that failed - it was trying to define a font
with an encoding vector only 128 entries long.  If you meet this problem
you can fix it if you are prepared to edit the PostScript file.
vector only 128 entries long.

Debugging
---------

If your PostScript program is failing, if you know a little PostScript
is usually quit easy to find out what is going wrong.  In fact if you are
developing PostScript code Post is a much better debugging tool than your
average PostScript printer.

It is normally best to use interactive mode for debugging.  Start up Post
on the screen and pick the "interactive" option.  Then run you program
from the interactive window.

To run your program, type:

    > (filename.ps) run

If it fails with an error message, you can then type interactive commands
to see what happened.

To display the contents of the operand stack, type:

    > pstack

To display the contents of the execution stack at the point of the last
error, type:

    > $error /estack get ==

To display the contents of any named variable, type:

    > varname ==

To see what is in the current dictionary, type:

    > currentdict {== ==} forall

If you have no idea how far your program got before it failed, try adding a
few lines "1001 ==" "1002 ==" "1003 ==" to trace how far it got.  As
PostScript files are pure ASCII text they are easy to inspect any modify
with any editor.