On Thursday 01 May 2003 20:54, Jimmy Wales wrote:
Another aspect of this, apart from disambiguation, is that identifying all motion pictures as (film) or (movie), even when there is no ambiguity, is that this "meta-data" could be used later on to extract a portion of our data for a movie history project or similar.
I thought about this possibility too and think it has lots of potential.
[snip]
It'd be nice to be able to do a query and pull out all biographies, for example, or all countries, or all movies, or all dog breeds. Not so nice that we should get all bogged down in debates about the perfect system, but nice enough that we should think about how conventions can help us move in that direction gracefully.
I agree. Categorizing the data, could also be useful for automatically generating overview lists like [[List of movies]], [[List of mathematicans]] etc. which are often incomplete and hard to maintain.
Perhaps the "meta-data" could be added in a similar way as it is done for the international webpages. For instance on the page for "Javan Tiger" one would add [[category:tiger]] in the very beginning. This information could be expanded in a "wordnet" like way, e.g. see
http://www.cogsci.princeton.edu/cgi-bin/webwn1.7.1?stage=2&word=tiger&am...
such that the "Javan Tiger" would automatically appear in the "list of mammals".
For what it is worth,
Marco