RE: [Axiom-developer] Boot

[Bill Page]

On October 2, 2006 9:17 AM Juergen Weiss wrote:
> ...
> I think, at some time, there was the idea to get rid of the
> old SPAD compiler (which contains a parser for boot, spad and
> the interpreter language). 
> 
> Therefore a new shoe boot parser in the boot directory was
> written, and a new parser for the interpreter (which is in
> use). The new compiler never happened and finally aldor (a#)
> was developed. 
> 
> So we have two parsers for boot and two parsers for the
> interpreter language.
>

So we can eliminate Shoe and also the old interpreter parser,
and everything will still work properly, right?
 
> There is actually a grammar file, which kind of defines the old
> parser language: fnewmeta.meta

Yes. Another argument for basing future work on boot on the
original boot version - at least there is some documentation
of the grammar.

> 
> About the boot ``language''. We had this discussion before: it
> really is not a new language at all, just another syntax for
> lisp. It has really neat features for destructuring and pattern
> matching. I would assume that anyone who knows lisp get hold of
> the syntax in an hour or two.

I am not sure why anyone would be afraid to call Boot a new
language although I agree that it's semantics are essentially
the same as Lisp (actually just a subset of Lisp).

> Bill Page and I argued in favor of boot for the time being, Tim
> Daly in rewriting everything in lisp. I just think, that as long
> as lists are the main data structure for the interpreter and the
> compiler, the boot syntax is really handy.

I also think that understanding that there is a close connection
between Boot and SPAD is an argument in retaining Boot. There
is a strong tendency for a new system to look like the old system
in which it is written. We definitely do not want SPAD to look
like Lisp to the mathematicians who write new algebra code. Boot
provides a natural transition from one environment to the other.

> A total rewrite in lisp with defstruct et al. of the whole system
> is too ambitious for the number of people contributing. Besides,
> for a total rewrite one should consider other alternatives as
> well (aldor, ...).

I still wish for this but I am afraid that is controlled by
people with a different agenda.

>...
> Gaby wrote: 
> > 
> > The more I read the interpreter code, the more I become 
> > convinced that Boot should stay.  It simplifies many lots
> > of logic; the benefits of removing it does not seem to me
> > to outweight the cost and the result.
> > 

It sounds almost like we should take a "vote" :-), although
open source is not organized on democratic principles as such.
It would be really nice if we could get consensus to proceed
in this manner.

Regards,
Bill Page.