[Cs22800] UC Free Software?

Mike O'Donnell odonnell at cs.uchicago.edu
Tue Dec 3 21:35:03 CST 2002


I'd love to have a good UC free software project. There are lots of
possibilities, and we just need to pick one. If we want to be the
center of the project, then we need to think a bit about
sustainability.

1. Open Network Handles, which I already suggested to Ben. I've been
writing about the idea

	http://people.cs.uchicago.edu/~odonnell/Citizen/Network_Identifiers/

The drafts are mostly pretty rough, but "Open Network Handles
Implemented in DNS" suggests an implementation based on DNSSEC.

2. Open Privacy Initiative projects. See
http://www.openprivacy.org/. Their idea of "Nyms" is technically about
the same as my proposal for self-assigned handles. But they take it in
a different direction, and work on identity/reputation management.

3. Loris/Fossa. See http://sourceforge.net/projects/loris/. Loris is a
library for sound manipulation, particularly appropriate for morphing
sounds (e.g., interpolating between your voice and a lion's
roar). Fossa is a GUI front-end for the most fun Loris
functionality. It was done pretty well as a masters thesis that I
supervised, but the creator graduated and it's stuck between alpha &
beta levels. There's also potential for some interesting extensions to
the functions in the Loris library, but that's tangled up with my
research on sound modelling---it's not just a programming problem.

4. ML/Moby/DrScheme projects for the programming languages faculty in
the department (Dave MacQueen, John Reppy, Robby Findler). How about
adapting the DrScheme interface to ML, or Haskell?

5. SWIG and/or multilanguage debugging projects for Dave Beazley.

6. I'm not sure what Anne Rogers is up to now. She's generally
interested in real-time data gathering. There might be a nifty project
in there.

7. Work up a kit for practical 3D visualization with interactive
motion. The basics are out there, but for some reason they never seem
to quite ring the bell. You could probably come up with something
pretty snazzy by integrating Geomview with the Gimp.

8. SKIlift. This is an interactive teaching tool showing derivations
in the Combinator Calculus (if lambda calculus is the assembly
language of mathematics, combinator calculus is the underlying machine
language). I have a rather crappy STK prototype at

http://people.cs.uchicago.edu/~odonnell/Computer_Geek/SKIlift/skilift.stk

I used it to provide an interactive movie version of the proof of the
recursion theorem for mathematically naive students. It's part of a
general idea of trying to present mathematical ideas to the
mathematically naive. It seems wrong that anybody can enjoy ballet
without being able to dance, but pretty much nobody can understand any
mathematics without being able to do mathematics. In at least some
cases, the right sensual presentation of the actual mathematical
content might do the trick. Along the same lines, I wanted to provide
some interactive visual presentations of primitive models of
computation, such as the 3-counter machine (three buckets with
pebbles---you can add and remove one pebble at a time, and check
whether a bucket is empty). Tarski's world and Turing's world provide
other ideas. They are proprietary Mac programs that are dying out. The
good ideas in them might survive if ported to a free project.

n+1. Continue any of the projects that you've already started on.

I started writing up an ideological description of a free software
group at http://people.cs.uchicago.edu/~odonnell/Citizen/PSST/, but I
never posted it publicly. If the basic idea is attractive, we could
expand on that. It's content-free: just a way of presenting the
work. But if people find it attractive, it could pull in more coders
and designers, and eventually some funding.

Mike O'D.



More information about the CS22800 mailing list