[Gsll-devel] Common approach to scientific/mathematical/engineering software

[A.J. Rossini]

On Mon, Jul 6, 2009 at 2:54 AM, Thomas M. Hermann<tmh.public at gmail.com> wrote:

> Greetings GSLL-DEVEL,
>
> The library I've been working on is basically a native CL reference
> implementation of a linear algebra API. It currently approximates the
> functionality of the BLAS. My intent is to layer it on top of lower
> level libraries such as GSLL once there is a certain amount of agreement
> on the interface. I have both working code and matching documentation.
>
> I can commit polishing up a first draft of the software and
> documentation for release when Liam gets back. We'll figure out where to
> post it then.

That would be great - especially if it sacrificed a bit of flexibility
for portability (Cyrus' CLEM is a nice CL library, but it's SBCL only,
which is a killer for what I want to do.

Would be great to be able to eliminate my current headache, which is
CL-BLAPACK (the headache isn't from the CL code per-say, but figuring
out how to use it reliably and automatically on a non-Debian system).

Liam - WRT to R, the pointer to Cyrus' CLSR is more for using his code
to figure out how to do it seamlessly, which he does better than I've
currently done, but using a few SBCL-isms, where as I'm a bit more of
a stickler for portable code (over optimization), trying to keep
things CFFI and Microsoft possible, if not being overly friendly
towards the latter.

Anyway, this month is "R-like Dataframes in CL" month for me, got to
talk about this hobby of mine in less than 2 weeks in Copenhagen, at a
conference on Statistical Computing (DSC 2009).  Will send a pointer
to the list when my slide deck (PDF) is ready.


best,
-tony

blindglobe at gmail.com
Muttenz, Switzerland.
"Commit early,commit often, and commit in a repository from which we
can easily roll-back your mistakes" (AJR, 4Jan05).

Drink Coffee:  Do stupid things faster with more energy!