Re: [Axiom-developer] Lisp portability

[C Y]

If I understand correctly, the issue is that various Lisp
implementations do not provide either a common interface to operating
system functionality, or a good way to provide those features if not
already present.

This is pretty hard to argue against, but unfortunately getting free
lisp implementations to agree to do something all the same way is
usually an uphill battle.  Sometimes this actually is because they have
different ideas about how to do things, but the INTERFACE level (api,
exported function names, whatever) I think can and should be much more
uniform.

An additional difficulty is that lisp programmers tend to solve
problems for themselves, and often times view these solutions in much
the same way most programmers might view a shell script written to do
some mundane job - "that was easy enough for anyone to do that it's not
worth turning into a library".  This leads to situations where many
individual developers have their own preferred solutions to problems
and (understandably) are not interested in standardizing on a different
way.  ASDF seems to run into this sometimes, with the instructions
being to customize ASDF for a particular situation rather than adding
the solution to the problem to ASDF proper.  Personally I doubt most
such problems are sufficiently unique to justify a one-off not being
folded back into the main asdf system, but apparently it's up for
debate.

My long term thought on this (I'll go ahead and admit I'm nuts to save
everyone the trouble of pointing it out) is to slowly convert the best
of the tools we use into literate documents and create the supporting
platform we need.  I'm slowly attempting to turn asdf.lisp into a
pamphlet file (figuring it out will be a learning experience) and even
if the asdf devs want nothing to do with it improvements can be added,
documented, etc. by us as work is done.  Eventually I would like to do
this with sbcl itself, cl-pdf, cl-typesetting, McCLIM, CLX, etc.  I
know that's probably an impossible goal, but it's proving (for me at
least) to be a good way to learn about a particular subject and may tie
in with Tim's long term vision.

Cheers,
CY

P.S.  Jay, would you be interested in pushing EAxiom forward some more?
 I never did get as far as the LaTeX parts of the original code you
developed but if there is more interest again it could motivate me to
push ahead.  Did EAxiom match how you were wanting to interact with
Axiom or are existing solutions sufficient?

--- Jay Belanger <address@hidden> wrote:

> 
> Gabriel Dos Reis <address@hidden> writes:
> ...
> > The point was more subttle than that.
> 
> Thanks; that's helpful.
> 
> Jay
> 
> 
> _______________________________________________
> Axiom-developer mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/axiom-developer
> 



       
____________________________________________________________________________________Choose
 the right car based on your needs.  Check out Yahoo! Autos new Car Finder tool.
http://autos.yahoo.com/carfinder/