home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Precision Software Appli…tions Silver Collection 4
/
Precision Software Applications Silver Collection Volume 4 (1993).iso
/
stats
/
chadyn.exe
/
READ.ME
< prev
next >
Wrap
Text File
|
1989-09-15
|
7KB
|
156 lines
SEPTEMBER 1989
Colleagues and students:
RECENT MODIFICATIONS IN DYNAMICS
Writing a program like DYNAMICS means making hundreds
of small decisions, and making hoardes of modifications to
ensure the parts work together smoothly and naturally.
One problem is to set up menus so that the user can find the
desired command, and there are many commands, so the less
essential commands should be buried more deaply in the
hierarchy of submenus. A continuing problem will be to make
sure the UNIX/X11 version is as easy to use as the MS-DOS
version.
Since the January 1988 version, the speed of the
numerically intensive parts has been sped up by a factor
of five due to a better compiler, and the program adapts
to the resolution appropriate for range of screens, and I
have included color graphics, and have found ways to make
color graphics useful.
To make full use of 16 colors on EGA or VGA screens,
you must have 640 KB memory, though the program should
run with 320 KB. See problem 1 below. The pictures can be
stored in color on disk, usually in less space than the
monocrome version reqired.
The menus have been reworked with the goal of
putting simpler information in front of the user by user
more menus. As a result some of the documentation is
inaccurate in the description of the menus. The command
descriptions are however hopefully correct. If a command
that you want is not in the menu, then try it anyway. It
is probably still there.
Documentation has been updated drastically, and that
process has not been finished. The disk(s) you receive may
have a file called MANUAL.DOC which is the partially revised
version. The new version includes more examples. There are
details on how to change the maps and differential equations
(though you must have a Microsoft C compiler for MS-DOS
machines). There is a file UNIX.DOC which desribes how
to proceed if you have a UNIX/X11 workstation.
The stable and unstable manifold routines
(introduced mainly in the summer of 1988) have
been made more robust as a result of a great deal of
detailed study. The straddle orbit routines have been
revised about once a month since H. Nusse and I now have
algorithms that are both efficient and (in nice cases)
rigorously justifiable. These stable and unstable routines
and straddle routines and basin routines are quite basic
to dynamics would be reasonable material for sophomore
dynamics courses.
In the past couple months, I have included routines
for following periodic orbits as parameters are varied; this
should be one of the more useful features of the program.
Currently it works only when the system has 2 dimensional
phase space, but there is no real difficulty in upgrading
the code so that it works in n dimensions.
There is now a variable step fifth order Runge-Kutta
solver for situations where it is important to control the
local error in solving the differential equations.
There is more ability to let the axes be parameters
of the system. For example, both coordinates of the Mandelbrot
set are parameters, and other than that, it is just an
application of the basin routines. You can use the small cross
to select a box and create a blowup, just as if the axes were
coordinates of y[].
UNIX COMPUTERS
Eric Kostelich has put a great deal of effort to
make the source code compatible with UNIX. The UNIX
version still has some minor rough spots. It has not been
thoroughly tested. It is set up to work in two windows,
one a text window for commands and one a graphics window.
The mouse is used only for pointing to a window. It must
point to the graphics window to recognize the keyboard.
See file UNIX.DOC.
KNOWN BUGS AND PROBLEMS
The program uses 568KB memory to provide
full color graphics. If your computer has lots of drivers
or memory resident programs installed, even 640 KB may not
be enough. While the program checks to see how
much memory is available, the operating system seems to
misinform the program in some cases, not detecting these
extra files in memory. If the screen refresh
(interrupt 'r') is putting lots of irrelevant garbage on
the screen, your problem may be just this situation. This
problem can be cured (in those cases found so
far) by using a different autoexec.bat file and possibly a
different config.sys for rebooting the computer so as to
eliminate some of the files and routines that are
ordinarily kept in memory.
GOALS FOR THE PROGRAM
The goal is not to finish the program but to have
a program that can be extended by the users. Since it is
constantly undergoing change, bugs can creep into places
that were bug free. Please report all problems with as
much detail as possible, including the detailed
circumstances. I would also like recommendations for
changes. I expect it to be published somewhere, hopefully
by September 1991. (at low cost). People getting this
version are acting as testors, and their comments would be
helpful, though please do not expect a direct response.
Computer programs for studying nonlinear dynamics
have been advertised for $600. I view this program
differently, as a package of ideas that I want to make
available to the scientific community, to as many students
and faculty and scientists as possible. I cannot give help
in individual cases. I view this package of programs and
documentation as the equivalent of a research publication
and I would like to request that papers that use it should
acknowledge it as a publication. Others will then learn
about it and will be able to benefit from its use.
Acknowledgements to the program should be in the the
references section and referred to as a book would be. For
example,
J.A. Yorke, DYNAMICS, A PROGRAM FOR IBM PC CLONES,
University of Maryland, College Park, 1989.
Several scientists have told me that it is stupid
to make the source code available, but I hope that they
miss the point. I will permit parts of the source code to
be included free of charge in other programs provided
permission is obtained in writing from me and provided
acceptable acknowledgements of authorship and copywrites
are included in the new program. I also look foreward to
others writing modules for inclusion in the program, with
acknowlegements of course. In the next version more
documentation will be available for those who want to
adapt the program to very different kinds of projects. For
the moment, peruse all the .DOC files included.
James A. Yorke