Development environment, was RE: more about Maxima + ECL

 
I would like a system in which all of Maxima can be read in and run
interpreted or compiled
and for which 
to_lisp();

(compile-file  <whatever.lisp>)
(load <whatever.o> )     or    whatever.fasl etc...
(to-maxima)

can be used to incorporate any corrections to the whatever.lisp file in the
system.

This is how Macsyma used to work. In fact, the first command in the
initialization
file for Macsyma was a file that read in all the newly compiled patches. By
so doing,
the system could be fixed in-between major releases, and for developers,
anyone can test a system with patches, compiled,
without doing anything more than a compile and load.

It is apparently the case that the run-time Maxima (GCL/windows) is no
longer compatible with
the compile-time environment used to compile Maxima.  

Unless I am mistaken in this, it appears to be a massive step backward in
system testing and development; in some sense a retreat from the notion of
interactive system development starting in the late 1960s  into the 1960's
batch development model of languages like C and Fortran.  Admittedly this is
"prettied up" by CVS, make, defsystem.   But still batch.

Am I mistaken?
If I am not mistaken, can someone try to explain why this has happened?

(My evidence is that starting with a running maxima,  re-compiling and
re-loading simp.lisp either gives errors, or provides a system that is much
slower than the original GCL/windows/5.13).

Thanks
RJF
 
> -----Original Message-----
> From: maxima-bounces at math.utexas.edu 
> [mailto:maxima-bounces at math.utexas.edu] On Behalf Of Robert Dodier
> Sent: Sunday, April 27, 2008 7:08 PM
> To: maxima list; sage-devel at lists.sourceforge.net; 
> ecls-list at lists.sourceforge.net
> Subject: more about Maxima + ECL
> 
> Hello,
> 
> I've been looking at Maxima + ECL a little bit more and here 
> is a summary.
> I committed ECL-specific changes to patches-for-ecl-branch in CVS.
> It would be extremely helpful if someone else would try the 
> ECL branch.
> Be advised that you must use the CVS head version of ECL, not 
> the 0.9j release.
> By the way I am working on Linux.
> 
> (1) Maxima + ECL, interpreted, works OK although there are about 12
> unexpected test suite errors. Of the errors, I see 3 kinds.
> (1a) different results in floating point computations
> (1b) simplification problems -- i.e. equivalent but different 
> expressions
> (1c) a type-coercion problem -- failed to coerce an integer 
> to a complex
> number or something like that.
> 
....