Re: Diffuse2d.

[cgl]

This is a good discussion...

One of the main things we are trying to do with Swarm
is to provide a venue where researchers
program up their experiments using higher-level
building blocks than just straight C-code....

  This means building models from "primitives"
that are the basic experimental apparatus of
our software laboratories....

  If everybody is going to use the same experimental
objects, there is a genuine pressure to make sure
that those objects are well understood and correspond
to the analogous "object" in their mental models of
the system they are trying to understand. It also
means that as a community, we can debate publicly
the merits of the various "features" in different
algorithmic approaches to implementing building-block
objects, and, by including citations to the literature
and test-suites to be run on the implemented objects, we
promote better repeatability and verifiability of the
models we build using those objects.

  As I've said before, if we can work together to build
a library of standardized *software* experimental
apparatus, with known behaviors and calibration
procedures, we can greatly improve the robustness
of our experimental procedures when we are trying
to do science via simulation. And others can 
repeat, verify, and/or extend our work because
they have access to the same experimental apparatus
we used in the community library of apparatus.

  Scientific simulation has been plagued too long by
one-off special-purpose monolithic blocks of arcane
code that nobody but the authors can use, read,
repeat, or extend.....expecially when the graduate
student who wrote it for 4 years has gone off to
spend the rest of his life "skinning otters in
Alaska" (a quote attributed to me..?!).

 Anyway, the more basic experimental building-blocks
we can develop through public discussion and community
design and specification, the better off we all are.
One of the main points of Swarm is to provide a general
purpose venue/format within which to implement these
common tools....

  That's the idea anyway - we need more of these discussions
and more contributions of basic experiment objects within
and from the simulation community...

  So, keep up the good discussion!

Chris Langton