[Axiom-mail] Re: Axiom, Rosetta and Quantian

[Dirk Eddelbuettel]

On Fri, Mar 12, 2004 at 11:10:16PM -0500, root wrote:
> Given that this is open source and "advocacy is volunteering" I'm
> willing to do what I can to make this useful. So my current thinking is:
> 
> (a) install Quantian (I just finished downloading the CD)
> (b) figure out how to modify Knoppix menus
> (c) add a math submenu with the various packages
> (d) get Axiom to run
> (e) add Axiom to the menus
> (f) figure out how to burn a live-CD
> (g)
> (h) profit!

The menus are often installed by the corresponding packages. So if the
Debian package of axiom that Camm provides has a menu entry, we don't need
to do anything more.  E.g., once you boot Quantian, you will find menus for
gap, yacas, maxima, ... and I didn't have to add those manually. "It just
happens" under Debian (if everybody does their job right).

That said, and in the spirit of the earlier point of 'some editorial
guidance', we way want to add a new layer on top of the existing one, and
for that we'd need a bit of work.  You can't test that approach with Knoppix
or Quantian on cdrom, though, as these are read-only.  If you use the
knoppix-hdinstall feature to install to a hd partition, then you're off to
the races.

> I'll work on (a) and (b) this weekend.
> 
> >> Can you point me at some info about how you build one of these CDs?
> >> I'll try to create one that has the proposed intro script.
> >
> >There is a document in the HOWTO section of the Quantian pages. While it
> >comes from an older Quantian version, all the individual steps are still the
> >same.  You need a loop kernel module to mount the iso, copy its conntent
> >out, mount the compressed KNOPPIX iso image with the cloop module (source
> >for cloop in Debian at at Knopper's site), write it out -- tehn chroot into
> >the expanded 'flat' filesytems, and at the end reverse the process by
> >building a compressed iso inside the outer iso.  
> >
> >More (and more comprehensible :) info is a knoppix.net.
> 
> As an interesting aside, one of my religious beliefs about this whole
> Computational Mathematics area is that we have to start combining the
> research papers with the code. The first cut I've made at this is to
> use Knuth's literate programming tools (actually, Norman Ramsey's 
> noweb version). I believe that unless we keep the research with the
> code we won't be able to maintain and modify these systems. 

You'd get on well with Tony Rossini who is also a huge fan of lit.prog. and
noweb :)

R (www.r-project.org) has a most useful extension called Sweave you may like
(do some casual googling). Works in Quantian as shipped.

> What's that got to do with Quantian? Two things are related.
> 
> Dr. Traverso at the University of Pisa in Italy is trying to create
> a Journal based on a similar idea where you submit papers that include
> the source code. Quantian would be an ideal way to "provide a base"
> for people to execute the embedded code. Ideally you could just 
> "drag and drop" the paper onto one of the systems and run the code but
> that is a long way in the future.
> 
> ISSAC 2004 is trying to publish its proceedings on CD this year.  If
> we get a good distribution of Quantian with a wide range of systems
> available we could make the case for distributing the CD at the
> conference and suggesting that ISSAC 2005 accept some papers built
> around the systems on the CD, especially those that show methods of
> combining the research with the implementation.

I'd be up for that, to the extent that I can help make Quantian a base from
which we could build, say, a 'pure CAS' flavour of Quantian by adding all
the CAS system, and tossing out some of the visualisers or other large
system in order to satisfy the 700mb compressed constraint.

Dirk

-- 
The relationship between the computed price and reality is as yet unknown.  
                                             -- From the pac(8) manual page