axiom-developer
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Axiom-developer] Community


From: C Y
Subject: Re: [Axiom-developer] Community
Date: Sat, 19 May 2007 21:54:57 -0700 (PDT)

--- Gabriel Dos Reis <address@hidden> wrote:

> I don't care replacing the Autoconf-based build system with something
> better; and if your system is better I'll lobby to use it.

My guess is the optimal system would be something similar to what
Maxima does - let the lisp asdf system handle the parts that can be
done in lisp, and the autotools system handle the parts outside of
lisp.  This is particularly true if we want to include and build gcl. 
Maxima's core parts (i.e. the lisp based code) can be built using only
lisp, and this is very useful in some development situations, but the
normal method is ./configure --enable-*lisp*, make, make install.  I
THINK this triggers the lisp system for that part of the build, which
must be done to bootstrap in any case.  Lisp build systems with
external triggers appropriate to the environment allow an encapsulation
of large parts of the build logic in a system that is insensitive to
system changes.

Maxima does this; I'm not sure about other applications of the
technique.
 
> I however have fundamental objections to any move to burry Axiom
> deeper into a second class build system, and second class galaxy.

I may be alone in this, but I don't view Lisp as "second class."  It is
a language with a long usage history, particularly in the domain of
problem we happen to be working on.  Someone on the list described it
as "assembly code" - that's actually not a bad thing in some
situations, and for others (like SPAD for mathematics) Lisp is
excellent for implementing languages to meet specific problem domains
(I believe Paradigms of AI Programming, one of the major lisp
references, illustrates that.)

As for ASDF, it is far and away the most widely used build system in
the Lisp world today - most modern libraries come with an asdf build
definition file.  Conceptually it inherits a lot from defsystem, but it
modernizes some of the design decisions and doesn't try to support
non-ANSI lisps.  The whole of the ASDF source code is 19 pages. 
Extensions needed for pamphlet files I'm not sure of yet, but those
will be part of Axiom and documented in literate fashion.

> If we ever want to attract more people to Axiom, the least thing we
> want is to paint ourselves into an exoteric ghetto. 

Of course, but I don't think Lisp qualifies as a ghetto.  It has a very
long track record and has stood the test of time quite well, in my
opinion.

> Please give a thought to why we don't seem to attract new blood to
> Axiom.  It is not that it is that difficult.  I've seen new blood in
> the more challenging and difficult GCC system. 

GCC is very general purpose (and widely depended upon), and equally
important perhaps it is very widely known.  Without GCC free software
as we know it wouldn't exist.  That attracts attention.

Right now, Axiom runs by default on non-ANSI lisp.  THAT version of
lisp is a ghetto, and is why I view moving to ANSI common lisp as a
step of major importance.  Dealing with non-ANSI lisp issues is not
going to attract many people.  Any old, non-standard source code has
similar issues.  Maxima gets slightly more activity than we do, but
they had a few years head start and are more portable at the present
time.  Even there, I think most of the active Maxima developers are
familiar names.  It would be interesting to look at number of new
developers acquired by more specialized open source projects over time,
to see what a common acquisition rate is.  (BRL-CAD might be a good
comparative example.)

SAGE is attracting a lot of attention, but if I understand it correctly
it's a meta-system that combines other systems to get a more powerful
environment.  That approach is quick (relatively speaking) and you get
a lot of bang for the buck programming wise.  That combination attracts
people.  Axiom's goals are somewhat different, which is one of the
things that attracts me to the project.

Shaking our reputation as "the hard CAS" would probably do the most
good as far as attracting people, but I'm not sure how practical that
is.

Cheers,
CY


 
____________________________________________________________________________________
The fish are biting. 
Get more visitors on your site using Yahoo! Search Marketing.
http://searchmarketing.yahoo.com/arp/sponsoredsearch_v2.php




reply via email to

[Prev in Thread] Current Thread [Next in Thread]