[Axiom-developer] Aldor files -- was: interval.spad INTRVL

[Weiss, Juergen]

> This reminds me to ask about the other ALDOR files that are
> included in the build but apparently not compiled. Are you
> planning that these will at some point be compiled by ALDOR?
> (which is now a separate non-BSD licensed? package) How
> difficult is it for a user to encorporate ALDOR if they so
> wished? There would likely be a considerable speed and memory
> advantage for certain routines. Right?

The ALDOR files besides interval.spad seem to be interesting
only to users of the link to the NAG mathematical library. At
least there are no SPAD files, which depend on them.

ALDOR was able to genereate LISP or C code. I don't know if this
is still the case. You could load the LISP code into AXIOM.
This should still be possible, though probably untested. I would not
expect any performance difference.

I was never directly connected to the AXIOM project, but from
hearsay I concluded, that at the  beginning of the '90s,
there was an old algebra compiler (the compiler which is still
used) and a new one, which was buggy, largely undocumented
and kind of unfixable. Stephen Watt and Knut Wolf and maybe 
others started to write a new algebra compiler from scatch
in C -- ALDOR. The ALDOR compiler has slightly different
syntax and semantic. I think the algebra (SPAD) files where never
ported to the ALDOR compiler. 

> Another thing that has happened since Aldor separated from
> Axiom is that a very large part of the algegbra has been
> re-written and apparently improved. Do you think that in
> the medium to long term the Axiom and Aldor algebra libraries
> should be made compatible? How practical (and legal?) would
> it be for Axiom (at least as a option) to implement the
> same library. Is it difficult to convert to spad. Or is
> there any advantage if ALDOR can be easily linked with
> Axiom?

Manuel Bronstein, at that time a main contributor to the 
algebra (integration, ODE) told me once, that he prefered 
a noninteractive system for the implementation
of his algorithms. So I think he adopted ALDOR and 
developped quite a bit of algebra in that language. I never
had a look at the algebra, which is available for ALDOR.
So I do not really know of which parts of mathematics
it consists and who contributed them.

At that time the machines where much slower and since then 
there were some optimizations in AXIOM as well. So the
development cycle of algebra code on AXIOM was quite time
consuming. The old compiler has its weaknesses too -- 
this was the reason the write the new compiler and the
ALDOR compiler in the first place. 

I think the compiler history is kind of symptomatic of
AXIOM as a whole. We have two parsers, the old one 
still used with the old compiler, and a new one now
used for the interpreter with some extended features
(parsing of rules for example). And I think this is
similar in other parts of the system.

After some history we should have a look at the future.
In which direction are we going? 

A few thoughts:

Though the system should be open (that is modular) so that
it allows the addition or the replacement of parts by developpers
who want AXIOM as a testbed for new ideas, the core
system should contain only one parser, one compiler etc...
Having everything at least twice increases comlexity beyond
the point where we are able to manage the sources.

There are parts of the AXIOM interpreter (graphics, hypertext, etc)
which I would consider optional components. The core system 
should run on any common lisp system without any additional
functionality which is now provided by C code. For example
the core system does not need xdr functions. This would make
initial porting to for example Windows much easier. 

Juergen Weiss

Juergen Weiss     | Universitaet Mainz, Zentrum fuer Datenverarbeitung,
address@hidden| 55099 Mainz, Tel: +49(6131)39-26361, FAX:
+49(6131)39-26407