On Tue, 2002-09-17 at 12:48, Richard Fateman wrote:
> When everything is put together correctly, the way to run
> maxima using allegro may be simply
> alisp -I maxima.dxl
>
> where alisp = the actual .exe for allegro CL.
> If I understand something said here recently, this can be put
> under an icon.
The idea was to encapsulate that into a wrapper script so that once
you've chosen a lisp for maxima, everything else is transparent. I had a
working system for this, but it has some problems, the most important
one being that it really doesn't make sense under windows. I won't bore
everyone with the details, but the command "alisp -I maxima.dxl" is
already there, in some sense, in the script maxima-run-lisp.
Unfortunately, there are some problems that stand in the way of it being
useful right now. After we get through this 5.9.0 release, I will try to
clean things up so that
1) We have a usable wrapper under windows.
2) Using an unsupported lisp is more transparent.
3) ACL is explicitly supported by the system.
> personally, I'd prefer that an emacs-like front end
> utter this command, and thus for me the startup would look like
> "start up emacs" perhaps by clicking an icon
> m-x run-maxima perhaps automatically by a special emacs+maxima icon.
That sounds like a good idea for the windows packaging people.
> the alisp -I maxima.dxl works for me right now, except I
> haven't looked at putting in plot etc.
Very good.
> The allegro maxima build uses the same defsys, and (other than requiring
> certain bugs to be fixed or conditionalized in the source), requires
> rather little. A half page of allegro-specific code to use its "dump"
> system.
I owe you mail on that topic from quite some time ago. Sorry. Let's try
to get those modifications into 5.9.1.
> But there are other issues like plot, interaction with front ends,
> which no one here has looked at. On the other hand, we have looked
> at a multi-threaded maxima, certainly possible in Allegro. Also
> a more elaborate alternative to simple socket code, like an http
> server. This latter stuff should be doable in other lisps as well.
That all sounds very interesting. Do you have anything particular in
mind for multi-threading?
--Jim