I’ve been thinking a lot about blogs and wikis. HTML seems so ill-suited to dense intertwinglings of thoughts. Ok for simple document layout, but the chunking/scoping seems at the wrong level for anything very fine-grained. I want to be able to pick this site up by a concept, shake it and see how everything else falls into place around it. Some approximation of that would be possible with heavy metadata and hyperbolic trees, but I feel like the map and the document should be the same thing. At the very least I want those visible interconnections within documents that Ted Nelson mocked up for Xanadu all those years ago…
have been busy attempting coding a riff on markov chains for multimedia. extracting inherent metadata and manually applying layers of it to dynamically create hyperbolic trees and hyperlinks (using statistical queues as with markov) … and then “regenerating” some media. basically, i get the feeling that unless your “picture” (=the algorithm behind its generation) has the balls to make some unlikely associations (guesses? “random” matches? …. or what I’ve settled on, mutations) then it’s probably not going to go much beyond a linear reading/hyperlinked association tedium.
have called it evolver. Has associated components:
evolver the main interface
involver a media input method
revolver a media output method, a viewer
devolver a media decoder
envolver a media encoder
the way things snap together is quite vague at this stage.
v0.1 should be done this summer. hmmm. tagline: “i’ll believe it when i see it”