Re: CAS funding for NSF (draft version, 5-27-2007)

On Jul 8, 6:44 am, Jaap Spies <j.sp...@xxxxxxxxx> writes:

JS> Maybe not for 'most people' but
JS> how trustworthy is a commercial CAS?
JS> Ask Vladimir!

Alas, as of 2007, things look BLUE.

A typical commercial CAS bears *at least* dozens
of thousands distinct defects. An amazing number
of regression defects.

Only Derive looks better. There are only thousands
distinct defects in Derive 6.1. Surprisingly,
there is an amazingly small amount of regression
defects in Derive, throughout all its history.

JS> This is where Vladimir Bondarenko fits in.
JS> Instead of finding bugs in commercial software
JS> he could be fixing bugs and solving problems
JS> in the open source alternatives.

Cyber Tester considers open source CASs, especially,
alphabetically, AXIOM, MAXIMA, and SAGE as a really
important part of pushing Mathematics worldwide via
making CAS maximally available and truly efficient.

We are going to address large scale open source CASs
defect identification as soon as we have more resource.

Best wishes,

Vladimir Bondarenko

VM and GEMM architect
Co-founder, CEO, Mathematical Director

http://www.cybertester.com/ Cyber Tester, LLC
http://maple.bug-list.org/ Maple Bugs Encyclopaedia
http://www.CAS-testing.org/ CAS Testing


On Jul 8, 6:44 am, Jaap Spies <j.sp...@xxxxxxxxx> wrote:
rjf wrote:
On Jul 6, 10:26 pm, Vladimir Bondarenko <v...@xxxxxxxxxxxxxxx> wrote:>

I've commented extensively on this white paper, privately.
But in response to VB's note

First of all I think it is not fair to comment on this two
key points from the first draft.

See:http://sage.math.washington.edu/home/wdj/research/oscas-nsf-white-pap...
for more details.

Key Points:

1. Funding research that critically depends on closed
source software is funding non-rigorous mathematics.

Proving that open-source symbolic math software is "correct" is quite
hard, and for most people who use a CAS, not of particular interest as
a research topic.

Maybe not for 'most people' but how trustworthy is a commercial CAS?
Ask Vladimir!



2. Funding commercial software via grant money that pays
license fees while minimizing support for open source
software is unsound.

I think that the NSF would view unfavorably a proposal to develop new
open-source software that merely duplicates existing commercial
software. NSF is supposed to fund research.

Here you miss the point. We need independent proof. The proof of a theorem
is published. Everyone can check it. This is not the case in close sourced
mathematical software. This is were Vladimir Bondarenko fits in. In stead
of finding bugs in commercial software he could be fixing bugs and
solving problems in the open source alternatives.

And as I understand it the NFS funds all kind of educational programs, not
only research.

Jaap


.

Quantcast