RE: [Axiom-developer] opinion

[Mike Thomas]

Hi all.

I would like to butt in here as I believe that there is an attitude barrier
which needs to be overcome - not an attack on Cliff's hard working and
enthusiastic attitude, but rather on a misapprehension shared by Unix
oriented programmers in general, I think.

Here is the misapprehension:

| Developing free software on Windows is a) very difficult

My personal opinion is that there is nothing inherently difficult about
Windows programming compared with Unix.

For projects like Axiom, Maxima, GRASS, GCL, CMUCL etc in which the software
has roots in the Unix/X11 world, it is the conversion of those projects onto
the different Windows (or the old MacOS) model which is tricky.

Catching up with many person years of (often formerly publicly or
commercially funded) development on Unix written with older software
engineering practices is a major contributing factor, not to mention the
esoteric world of Common Lisp in those cases we are discussing here.  (I
personally reckon that the GCL C and Maxima Lisp source trees are the most
unreadable I have ever seen. Axiom source (not the Latex pamphlets which are
horrible for programming purposes) I find very readable on the other hand.)

An excellent example of the difficulty of crossing the programming paradigm
boundary in the opposite direction is the trend of bringing music production
software into the Linux/*BSD arena - a slow and difficult process on those
platforms in a field traditionally dominated by MacOS, Windows and Atari
TOS.  The Agnula/DeMuDi people are doing a great job with great progress but
lag behind the software offered on Windows/Mac.


| and b) very
| frustrating.  People like Mike and David have a very special skill set
| that is difficult to duplicate and takes a special kind of devotion.

(We're actually "on a mission from God" - note the quotes before taking this
seriously.)

| >From the Maxima experience, getting things working on Windows is always
| the most exasperating part of any release.  The bugs are easy to find
| but hard to fix - they vary between different versions of Windows, and
| even between different installs of the same version of Windows.

This variety is similar to Linux/*BSD/Solaris/IRIX/GNU library and tool
versions I believe.

| The
| port of the gnu toolset to Windows (mingw and friends?) is the only way
| to generate a stand alone binary for Windows using only free software,
| and I think it is fair to say those tools are still evolving.  When you
| add all this up, plus the average profile of your Linux/*BSD users (on
| a statistical basis) vs. your Windows users, the odds that a Linux/*BSD
| user will be up on at least the basics of software development, and
| have a well integrated toolset for developing already on his/her
| machine, are MUCH higher.  So Windows developers are rare as a
| percentage of all Windows users, free software Windows developers are
| rarer still, and the intersecting set of Windows free software
| developers interested in CAS work is downright tiny.

Agreed.

There is also a strangely branched attitude around on both sides of the
software divide which says "You Windows free software developers are [in
some sad way ethically unclean (this from the Linux world) | stupid for
doing free work (from the Windows/Solaris/IRIX side where _all_ programmers
I meet are commercial programmers)]."

I encountered the first attitude in person face to face from one of the
leading lights in the free software community (the only open source
developer I ever met in person) and the latter from almost every programning
colleague I have ever worked with.

| Eeep.  Now that I thought that out, I'll take this moment to say -
| THANK YOU, Windows developers!  And thank you Tim Daly (and Bill
| Schelter) for making all this possible in the first place.

Thanks Cliff.

Cheers

Mike Thomas.