Re: [Axiom-mail] Re: AXIOM available?

[root]

Arthur,

re: Bourbaki. I can't say it was intentional. I originally left the author
tag out of the files. At one point I wrote a script that did a mass change
of all of the pamphlet files to include the tag. I'll fix it before I
next upload the sources. 

re: Bourbaki. Is that an organization? I know it was the name used by a
group of anonymous mathematicians but was unaware that there was an
organization by that name. Rats. It seemed so appropriate to have the
default name be one that is recognized as a group of anonymous mathematicians.
Now I'm gonna have to get creative again. I could list my name by default
but that also seems inappropriate (and egotistical to boot). Sigh.

re: home directory of /home/axiomgnu/new
Again, another default decision. I have about 30 "axiom*" home directories
on my machine. The one that was most current when I was ready to do the
upload was in the axiomgnu directory. Thus, the chosen path. On savannah
I plan to do two things intentionally (lest you get the impression that
ALL decisions are thoughtless). The CVS tree is going under axiom and
the executable binary trees will go under spad. 

re: combining bits of GPL-licensed code. You joined seconds after a
3 day packet-storm discussion about GPL led me to create an axiom-legal
mailing list. I hate to say it but I can't even BEGIN to think about
GPL issues for another couple weeks. The whole discussion makes my 
hair hurt. But if I get to thinking about it again or you'd like to
raise any issues, please use the axiom-legal mailing list. All license
and legal discussions are banned from the developer list. 

re: CCL. The CCL code version is in plan to be used. I found it hard
to get started because I couldn't figure out the whole model you were
using, e.g. how to compile files, build and save images, etc. You can
see that I spent time in the pamphlet files trying to document how it
works. GCL, at least an axiom version, will eventually be subject to
the pamphlet file treatment also as soon as time permits. Pamphlet files
let me pass on my understanding to the next developer (so they don't
have to reverse engineer the code) and I've gotten religion about it.
As I revisit the CCL code, for example, to figure out what ONEP does
or how openmath should work, I'm documenting the code. If you want the
original code (or mods that I make) as lisp code you can just type:
   notangle foo.lisp.pamphlet >foo.lisp
and the original files return. I understand that you aren't likely
to use the literate programming style.

I would make the comment, however, that a key difference between open
source code and proprietary code is the notion of training. You wrote
the code and you maintain it so, clearly, there is no need to explain it.
When I wrote the Axiom code I didn't need to explain it either. However,
if I expect Axiom to live and be maintained as open source I need to 
pass on everything I know about the code, how to build it, modify it,
maintain it, and extend it. Literate programming is perfect for that.
Comments in the code are usually too terse and assume you understand
the code. External documents don't get written or, if they do, don't
get maintained. If you consider that Axiom, over the next 30 years,
is going to be maintained by a sea of developers then there has to
be a mechanism to pass along all kinds of information. Without it even
experienced lisp coders have a hard time. I've programmed in lisp since
college and I still had a hard time cracking CCL though now I have a
much better understanding. Literate programming only adds 2 tags to
a latex document so the learning curve is minimal.

re: congrats. Thanks. 

Are you going to be at ISSAC?

Tim
address@hidden
address@hidden