From Raph Levien's Diary:
Paul Snively has been blogging about proofs as well. I've skimmed his references, but so far am not sure how relevant they are.
Well, that's fair; I'm not sure myself. OBJ3 and OSCAR have both been used for theorem-proving and neither of them had been mentioned before, AFAIK, so I thought I'd point them out. I thought that Raph might appreciate OBJ3's modules and perhaps OSCAR's performance and flexible architecture.
Meanwhile, I continue to be interested in Mitch Kapor's new project. It might even be the thing that finally induces me to learn Python. However, I read the Slashdot thread, and while I generally find Slashdot's noise:signal ratio unacceptably high, I do have to echo one of the questions raised there, which is: if the team has been "thinking about the architecture" for a year, where are the architecture docs? Where are some proof-of-concept modules? Why does the design mailing list include the kicking around of such fundamental stuff as how Chandler will discover and communicate with other instances in a peer-to-peer fashion given that no one really seems sure how the Jabber infrastructure does P2P, or even whether Jabber is the right tool for the job? Why isn't there a discussion of how RDF will be used, what existing schemata might be useful, what new schemata might be necessary? A discussion of how RDF queries would be implemented, and what relationship this would have to Chandler's anticipated "automatic classification" features?
Still, I love the idea, and hopefully I'll have something to contribute in one fashion or another along the way. And I'm still interested in Raph's trust metric work, too—in fact, I see some overlap between the RDF work and the trust-metric work, about which hopefully I'll have more to write at a future date.
10:49:07 PM 
|
|
