[Axiom-mail] Re: Axiom, Aldor and program proof.

[root]

> Martin Dunstan wrote:

> > Tim Daly wrote: 
> > to become an Aldor developer because I need Aldor running on the
> > Palm-pilot class of machines. I'm willing to do the work to port

> *grin* When I first read that sentence I wasn't entirely certain how Aldor
> would run on a Palm given the memory and storage requirements. After a bit
> of thought I realised that a Palm probably has more CPU power and RAM/disk
> than the Sun Sparc's I was using when I started on Aldor a few years ago!

I'm looking to get the newly announced but unavailable Sharp Zaurus
400Mhz, 128M, etc. It runs full linux on an Intel processor. Porting
should be fairly easy. Until that time I'm on a Palm which could be
made to run the code so I'm looking to retarget the back end of the
compiler. I'm trying to convince everyone on our project to use Aldor
as their development language as we're doing group theory work and
Aldor would be perfect. Plus I need to create a libPalm library for
Aldor that exposes the palm api and gui.

> > significant dent in the problem. This will take a few years, of course,
> > but I'm looking toward a 30 year horizon for Axiom so I'm not concerned.
> 
> Wow - a 30 year horizon must be very rare for any project.

Axiom's been around for 30 years or so already. I expect it will last
at least as long in its open source version(s). We've got several people
with various interests, like program proof, new front ends, better algebra,
new algebra areas, teaching, etc so things are moving along quite quickly
despite the fact that it still isn't available (but hopefully by year-end
for the algebra portion).

Given the current funding model we're able to take Axiom anywhere we
can find an interesting thing to do. And there are dozens of people
who have interesting ways to use a computer algebra system.  This is
unlikely to change except to spread so the 30 year horizon is not 
unreasonable. I still consider Axiom a work-in-progress research
project and I believe that research ought to have long term horizons,
otherwise it is just development. Of course, somebody else will be
"driving the bus" by then so that's the whole motivation behind the
documentation effort.

> > Actually, I'd prefer a TeX form of the document provided you were willing
> > to let me rework it into a pamphlet form.
> 
> Sure - attached is a tar.gz file. Under Linux you just need to type:

Thank you for sending the tex for your thesis. It makes it so much
easier to work from the original sources. Once I get to the algebra
level of the system I'll look at the details of how to work it into
a pamphlet.

> I think the camera that they want to use for this particular mission are
> 1024x1024 at 20Hz and 10 bpp. The visualisation simulator I've been 
> working on currently manages up to 1600x1200 at about 4Hz and 8bpp so I'm 
> not too far off real-time ;)

I used our camera to drive a robot that was arc-welding a seam between
large metal plates. Since everything was either BRIGHT WHITE or SCREAMINGLY
BRIGHT WHITE the 10bpp would have been overkill :-). I do envy you the
chance to work in vision though. It was a fun area when combined with
robotics work. If they need someone to repair it on Mars, call me.

Tim