[Axiom-developer] Of possible interest to Axiom.

[Stephen Wilson]

*,

I sent the following message to Tim.  Posting here to archive any discussion it
generates.

Unfortunately, these are still early days, and I am not yet ready to release
the code.  On the other hand, the design space is still fairly open, and I would
be thrilled to discuss the possibilities with anyone who takes interest.

Take care,
Steve

Stephen Wilson <address@hidden> writes:
> Hi Tim,
>
> I thought I would send you an email letting you know of some work that I am
> doing.  I do not know if this would interest you or the Axiom project so will
> try to keep this message short.
>
> As you may recall, I have an interest in compilers.  I wanted to improve SPAD
> and help rewrite it in CL.  The boot code almost killed me and so I started
> thinking about doing things from scratch.  That was about a year ago, I think.
> Since then I have been thinking about the problem and eventually made a 
> decision
> to implement a compiler for a "small" version of SPAD, and then use that as a
> staging ground for a more powerful system.
>
> The work to produce a small language and compiler is still ongoing, but quite 
> a
> bit of progress has been made.  The system is implemented as a set of lisp
> macros (for now).  An incomplete example:
>
>    (defdomain polynomial-ring ((R ring) (E ordered-monoid))
>       (:satisfies ordered-monoid (retractable-to S))
>       (:carrier rep (record (c R) (k R)))
>       ...)
>
> The system is in its infancy and is not ready for general consumption.  But 
> even
> now I am coming upon questions of design which would be influenced by the
> possibility that such work might be of benefit to Axiom.
>
> Although the above is not much to go on, I am curious about any thoughts you
> might have.  Although the language is currently small, it has well defined
> semantics, an ever cleaner implementation, lots of comments (could become
> literate), etc.  As it is lisp all the way down, with a rich ast (hierarchy of
> CLOS classes) representing domains and categories, I think there is a lot of
> potential (time will tell).
>
> Is this something of interest to Axiom?  If so, I would appreciate having a
> dialogue about the issues, on axiom-devel if you prefer, as it would help me
> develop a picture of where this effort is headed in the long term.
>
>
> Thanks for you time,
> Steve