Re: [Axiom-developer] Re: [Gcl-devel] gcl sockets

[root]

B,

You raise several issues in the same paragraph. Lets see if I can
answer them separately.

(1) your planned AXIOM plugin into apache is very ambitious, however, 
    wouldn't a small perl script be a sufficient first step? 

I've read the mod_lisp.c code. It isn't very complex as these things go.
It simply talks to a lisp process running elsewhere thru sockets. 
So I have to:

(a) get a lisp running that talks to sockets
(b) write an Axiom domain to talk to the server

Sockets are how Axiom talks to it's graphics process and I'm in the process
of recovering the graphics so it's pretty much the same problem. The first
step is to execute an S-expression over a socket which is what I'm working
on at the moment.


(2) ..........  Indeed a CGI something cannot
 handle multiple command in several HTML pages, but one can submit (using a
 form) a rather complex multiline command to AXIOM. 

Initially it is planned to handle single commands. However I don't see
any big problems to handle multiple commands. I can use the session
identifier as the name of an Axiom namespace so that multiple users
don't affect each other's results.


(3) ............ Including possibly  system commands however. 

Dylan mentioned this. The Axiom domain will filter commands to disallow
things that use system commands. 


(4) ....... Do you think that a normal web server has the
 potential to handle a possibly computational intense AXIOM task too?

In general I expect the web server to be barely registering in terms
of process time. I can't imagine tens of thousands of people trying
to do anything on the server at the same time. Since Axiom is running
in a different process (or possibly on a different machine) the web
server should be unaffected.

As for long computational tasks:

(a) initially I expect Axiom will only be running on one processor and
    used to develop documentation thru Page's wiki server.

(b) in the long term there is a sub-effort to put Axiom on a beowulf.
    I have access to a 128 node cluster with 2Ghz processors on each node.
    Of course getting Axiom running in small-grain parallel mode (as 
    opposed to an independent Axiom on each node) is a hard problem
    on its own. I've done some initial experiments with a matrix domain
    (parallel matrix multiply) but not much since then. I was hoping to
    have a student to work on this but that fell thru.

(c) If Axiom reaches THAT level of interest perhaps somebody might be
    interested in funding it :-)

t