I believe that many people are working on ways to index pictures, [link|http://www.springerlink.com/(n4fp4t45ji0sko45smmnum55)/app/home/contribution.asp?referrer=parent&backto=issue,7,7;journal,3,32;linkingpublicationresults,1:101757,1|movies], [link|http://www.music.mcgill.ca/~knopke/publications/icmc2004template.pdf|music], [link|http://john.cs.olemiss.edu/~rice/FindSounds.pdf|sounds], etc., based on the contents of the files rather than the filenames. With fast enough computers, and good enough algorithms, it probably won't be too long before bots can figure things like this out.

It's interesting though.

Cheers,
Scott.