axiom-developer
[Top][All Lists]
Advanced

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

[Axiom-developer] reqeust


From: root
Subject: [Axiom-developer] reqeust
Date: Fri, 11 Nov 2005 00:26:11 -0500

So far i've been able to set up a live CD that can be booted
and does not touch the hard drive (the KNOPPIX idea). 
(see sourceforge.net/projects/doyencd)

I have set it up so it starts up a webserver at boot time. 

I have set it up with a browser that contains a page with a text box.

The CD contains Axiom and the literate tools.

It starts a web page with an "expression box" Typing an expression in
the box and pressing enter will cause the expression to be evaluated
with the results returned to the page as pretty latex output. This is
essentially where mathaction is today.

Now I need to be able to do two things:

 1) make the web page look and act like an embedded shell

 2) drag and drop a file onto a web page and have it automatically
    "uploaded" (of course, it's all running on the same machine but
    don't assume that).

Can AJAX help with these two problems?



Essentially I want the following scene to work:

  1) I go to a conference with a bootable Doyen CD which contains the
     proceedings, axiom, the browser, literate tools, etc.

  2) everyone at the conference gets a copy of the CD

  3) a speaker gives a talk at the conference and, as part of the
     talk, indicates a URL where a literate copy of the paper lives

  4) the conference attendees boot the proceedings CD and visit the URL

  5) the conference attendees "drag and drop" the paper to their machine

  6) drag and drop causes the following actions:

     a) the paper gets the latex extracted (weave)
     b) latex is run on the paper
     c) the dvi or pdf output shows up in a browser tab

     d) the paper gets the code extracted (tangle)
     e) the code is compiled and loaded
     f) a tab opens which has a command prompt
     h) the user can interact with the code


Consider a simple paper that compares two sort strategies (bubble
and quick). The presentation shows a table of results:

          

          1 10 100 1000 10000 (# of random entries in list)

bubble
              (table of comparitive execution times)

quick

While the talk is ongoing the conference attendee can fetch the paper
(which contains the code) and see the actual computed results on their
local machine by running the code.




This is the "Doyen" concept, in essence. It will allow us to promote
literate programming at all of the computational science conferences.
Real code can be distributed and run and even updated in real time.

It will eventually be the case that presenting computational science
results without presenting the code will be like presenting a theorem
without presenting a proof. I'll believe the results in the table when
my system arrives at the same results.

Somehow it seems just a small step away from where we are but my 
web/browser/javascript/xml skills are lacking.

t




reply via email to

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