[Axiom-developer] Re: Doyen

[bill.page1]

"Tim Daly Jr." <address@hidden> wrote:

> On Wed, 2005-11-16 at 00:24 +0000, address@hidden
> wrote:
> 
> > Must people don't run Apache on their desktop configurations.
> 
> What?  I don't know these most people, but most of the people
> I know run Apache.

Oh, that's good. I thought that most Linux distributions considered
Apache a "server" application that was not installed by default in
a "desktop" configuration. But if more people are running Apache on
their desktop computers, I think that's fine.

> 
> > Perhaps for performance reasons on a network mod_lisp might be
> > preferred although from what I hear Araneida is supposed to be
> > pretty good. Is any other reason why Tim Daly Jr. recommends an
> > Apache solution?
> 
> I don't know exactly what you guys are doing, being CC'd in the
> middle of this and all,

Sorry, you were just on the Cc list to which I was replying
regarding the "Doyen project" about which I presumed you did have
some previous knowledge. We are really "doing" much more than that.
But anyway, I appreciate your reply.

> but I wouldn't think about performance until you have
> something working.  
> 
> Apache is solid, well-tested, and well-maintained.  For a
> hostile environment like the internet, it's an appropriate
> choice.

Agreed. In the case of Doyen, however this is an application
that is intended to run locally on the desktop - just another
gui application except for the choice of web browser as user
interface. It is usually not difficult to protect such a
application from the internet.

> The author of Araneida runs it behind Apache, both for his
> commercial sites, and for the cliki.

I think that is a good strategy. On the axiom-developer.org
server we also run Apache as the front-end with Zope behind it
for similar reasons.

> 
> TBNL provides a more traditional framework than Araneida.
> I think it's also slightly more comprehensive, and certainly
> better documented. ;)
> 
> > But cliki itself seems like a fairly neat and tidy wiki
> > implemented in lisp. If someone was motivated to add LaTeX
> > and Axiom integration to cliki then that might be quite
> > interesting.
> 
> FWIW, anybody who wants to play with cliki would probably
> have a lot more fun with cl-typesetting than LaTeX. :)

I've looked at cl-typesetting. It's nice but it doesn't get you
very far if your source documents are all in LaTeX and Axiom
generates LaTeX output. In AxiomUI for example it is clear that
jsmath on the browser gets you a lot further.

> 
> > Since http://www.cliki.net lists Maxima as a common lisp
> > application and GCL as a common lisp implementation and
> > Axiom runs on GCL it makes sense that cliki.net should also
> > include Axiom, so I have just added it. If you get a chance,
> > you might want to review what I wrote.
> 
> Axiom runs in gcl's special "archaic" mode; it's not a Common
> Lisp program yet.  But adding it to the cliki might help that
> process along. :)  Thanks for adding it.

Is Maxima compiled on GCL as a Common Lisp application?

I think upgrading the Axiom source to Common Lisp is something
that should be done with high priority. If I had the necessary
skill, I would probably be working on that right now.

> 
> > 
> > Wouldn't it be nice (hint, hint :) if someone decided to do
> > an SBCL port of Axiom?
> 
> I started on it, got a good chunk of the compiler compiled
> (which is a long way from working...), and couldn't maintain
> interest because (at least that part of) axiom's code is so
> awfully horrible. It seriously needs a rewrite, and I don't
> have the time.

I thought Axiom was supposed to be quite portable - having been
previous compiled on a long list of pre-historic lisps. From
what I understood, from Tim, just about the only thing
preventing it from be Common Lisp was some non-standard use
of packages. Ignoring any revulsion you might feel for working
on awfully horrible code, having worked on moving Axiom to SBCL,
is it your experience that the changes required to make Axiom
conform to the Common Lisp standard are much more than that?

> 
> > > "embedded command window" is also possible. it appears that 
> > > franz already offers this ability with a lisp prompt in a
> > > window. they appear to be using a java applet which opens a
> > > tcp connection to the host-based lisp process and emulates
> > > a terminal.
> > 
> > Very "old tech"... I don't see why you think this is relevant.
> 
> Actually, the web repl is only a year or two old, I think.  What
> could you possibly think is "old tech" about it?  What kind of a
> term is that, anyway?  Software is not seasonal.

It's a pejorative term. I know people still use terminal emulators,
but the point of Doyen is a graphical user interface. I don't like
the idea of slapping some console-style terminal emulation inside
a web browser and calling that progress.

But perhaps I completely mis-understood. What is a 'web repl'?

> If you don't understand the "old tech", you're just going to
> re-invent it, (probably in a bloated, embedded-in-a-web-browser,
> xml-compliant, and thoroughly unprincipaled way).

I agree. In fact, that was basically my point.

> 
> > If you want to configure Apache on axiom-developer.org with
> > mod_lisp I would be glad to help although I doubt that I will
> > be much inclined to start developing web applications in lisp
> > any time soon. Compared to Zope (which some people seem to
> > think is pretty complicated) all this lisp-based web stuff
> > seems completely over the top to me... ;)
> 
> Working on a web application written in Lisp might be a nice way
> to learn enough of it to work on Axiom.  Despite rumors to the
> contrary, web applications are often simpler than computer algebra
> software.
> 

With the proper tools, I think a web application (even in lisp)
should probably be somewhere between 100 to 1000 lines of code
so I don't know how much opportunity that would give me to learn
much more about lisp. :)

But really, my point is rather similar to yours at the beginning
of this message regarding Apach+mod_lisp. Zope works. It has
some very high level applications available off-the-shelf and
it required about 100 lines of python code to integrate with
Axiom - so why not use it?

Regards,
Bill Page.