RE: [Axiom-developer] Aldor and Lisp

[Page, Bill]

On Tuesday, October 18, 2005 2:25 PM Gabriel Dos Reis wrote:
> 
> Bill Page writes:
> 
> [...]
> 
> | Although Aldor is still not open source
> 
> Is there a hope that it will ever be?

I think there is still a good chance that Aldor will be licensed
for open source. I called Steven Watt again today. He wasn't there
but I left a detailed voice message. I told him that there have
been some discussions on the axiom-developer email list about
writing a new open source version of Aldor and that in my opinion
it would be much better if Aldor itself could be made available
under an open source license. I hope to hear from him soon.

I think it would help if some of the other subscribers to the
axiom-developer list were to contact Steven Watt and make a similar
argument. I understand that he is busy with other research (aren't
we all?) and I don't mean that we should harass him in anyway about
this, but I do think that it is important that he understands that
the availability of Aldor as open source matters to more than just
a few people. 

> If not, what are the odds of creating a freely available compiler
> that accepts Aldor programs?
> 

Unless a major workforce of several knowledgeable compiler writers
step forward who know lisp and have experience with SPAD in Axiom,
I have serious doubts that it will be possible to write a new
open source compiler that accepts Aldor programs. After all, Aldor
is a complex very high level language - not so different from
Haskell or Ocaml and it certainly took a major effort to implement
usable compilers for those languages.

An alternative that does exist (maybe) is to make incremental
improvements to Axiom's built-in SPAD compiler that would make
it more compatible with Aldor. In effect this would amount to
"back-porting" some of the improvements that where made to the
SPAD language when Aldor was originally implemented. The result
will be a compiler that works within Axiom to write better Axiom
library code. But it would not have the option of running as a
stand-alone compiler the way Aldor can now.

I still think however that even improving SPAD will not be easy.
It will require rather deep knowledge of the largely undocumented
legacy code that currently implements SPAD in Axiom. It is notable
I think, that the original Axiom developers chose to re-implement
SPAD as Aldor (written in C) rather than continue to make
improvements to SPAD.

Regards,
Bill Page.