Re: [Axiom-mail] lattice graph program

[root]

Martin Rubey wrote:
> is this what you have in mind?

> > I've been messing with XREF a little more, and using graphviz and xref
> > I made a chart which maps the risch.lisp functions, as well as the
> > RZMAC and RATMAC files.  I doubt it's useful for much but I thought I'd
> > upload it in case anyone is interested.  One that is definitely useless
> > is a joint mapping I did of all the rat*.lisp files.  It's way too
> > complex to be useful, but does give you an appreciation for just how
> > complex this program really is.  I uploaded the postscript files in
> > case anyone is curious:
> 
> > http://maxima.sourceforge.net/risch.ps
> > http://maxima.sourceforge.net/rat_files_chart.ps
> 
> > CY

I don't know but I'll check into that. Thanks.


> BTW: two questions:
> 
> 1. Is there an estimate when there will be a first axiom (alpha :-°) 
> release?

The first release that includes compiled algebra code will happen this
weekend. I've finished the remaining domains and, except for a few
that failed to compile, the whole interpreter/compiler/algebra code
will be available to developers. Once some developers have
successfully compiled it I'll turn my attention to releasing the same
codebase on savannah. I've been working on a private CVS so I can make
my initial mistakes there. Now that Axiom compiles algebra from
scratch it is becoming usable. Of course, "usable" is questionable as
it is nowhere near the version that NAG shipped. Much work needs to be
done both to fix up the algebra layer, to build the rest of the code,
to clean up the docs, etc. However, the algebra is the important part.

I've done almost no checking of the code at any level except that it
compiles and builds properly. There is a tremendous amount of work
just to remaining to check the code so don't assume that this is a
finished release in any sense.

> 
> 2. Why is axiom NOT part of the gnu project? Wouldn't you at least ask 
> for something similar as (from aldor)

The code was released under modified BSD. That makes it both free and
open source in the technical sense. The GPL license was not chosen because
of various issues. Some of the discussion about the license still exists
at the original axiom site (www.earthlink.net/~jgg964/axiom.html).

Aldor is under its own license. Contact www.aldor.org for details.
The Aldor software is a separate effort of which, so far, I'm not a
developer. 

> 
> If you distribute a modified form of either source or binary code (a) you
> must make the source form of these modification available to Aldor.org (b)
> you grant Aldor.org a royalty-free license to use, modify or redistribute
> your modifications without limitation, and (c) you represent that you are
> legally entitled to grant these rights and that you are not providing
> Aldor.org with any code that violates any law or breaches any contract.

I'm confused about the intent of the above parpagraph. Axiom source code
is (or will be) freely available to everyone. It isn't usable in the
current state as it is just a "box of parts". I'm automating the build
process so anyone can build a working system. This involves many (8, so far)
layers of code, some of it self-referencing and circular.

As the build procedures begin to work I'm releasing the code. Aldor
developers, as well as anyone else, are free to do anything with that
code including NOT making their changes available (per mBSD). I expect
to work closely with the Aldor group in the future. 

> 
> (if there was a choice, I would find it great if axiom would become gpl, 
> but presumably there are lots of legal issues I do not know about...)
> 
> Martin
> 

Axiom code is 30 years worth of work by about 100 people and went thru
two companies (IBM and NAG). Both have been extremely generous in open
sourcing the code. The choice of modified BSD makes the code as free as
it can be. 

Tim Daly
address@hidden