January 31st, 2003
|01:31 am - concat|
Figured for my own reference I'd pass this along. If you haven't seen it, this is surreal and interesting. And the switch ads seem to be a popular topic of parody (Linux, Rosen)
Another attempt at "Total Information Awareness" (not yet scheduled for preemption!) tomorrow. WRCT 88.3, 6pm. Which means I probably have to make my mom take the 28X (and given Friday afternoon Pittsburgh traffic...)
A somewhat amusing take on insta-wealth (on inheriting $2.7B++).
9. Open a savings account for college.
8. Buy a college.
|Date:||January 31st, 2003 08:35 am (UTC)|| |
The heirarchical filesystem is a really wonderful thing for programmers and websites. I think you expose yourself in your retraction.
Part of your resistance is a fear, I believe, of the performance penalty of amorphous data. If we ignore that (and you are enough of a database haxor to recognize how this penalty could be reduced), I think I agree with a bunch of it. Of course, I do have stuff directories.
In your final example, for instance: If you wanted to listen to Orb, you'd type "orb". I'd imagine for music this system would be even better than for normal data -- did you ever use Napster? It is much more intuitive, if you know sort of what you want, to be able to enter some of the parameters. An example: I like Elliot Smith's song "Needle in the Hay". I don't remember which of his albums it is on. I could either say "I want Elliot Smith's Needle in the Hay", or I could change into the Elliot Smith directory and hunt (manually or with a search) through his discography. Why don't I just not separate by CD, you say? Because whereever I make my division of bins, I lose either ease of use or information (what if I want to load up a particular album?) If the filesystem could read ID3 tags, none of this would be an issue.
I guess I'm just not organized like you are.