Started to think about Beagle indexing today and how you need to have textual data added to the file in order for Beagle to actually store some information on it (e.g. EXIF info in a JPG file). Wouldn’t it be cool if say Beagle could scan your Mp3’s, find out who the artist and album are from the Mp3’s data and then go out on the net and drag down the lyrics for that song and index that related to that file. Or if you have a ogg file on your Hard Drive from a meeting you had, Beagle could run the file through a speech to text translator and index the contents of the meeting. It would be cool if it also stored the time the keywords were said so if you searched for the string “Kill the president”, Best would show the ogg file with a timeline with sections highlighted and you could play just those sections.
Images and video would be a bit more difficult but it could work on the same concept. Beagle would scan the video file and if it found someone’s face it recognized from Evolutions Address book it could index that part of the file with a link to that person.
Then again, you might as well use that function to automatically tag images with group lists from Evolution to be used by Nautilus or FSpot. People who appear in your FSpot Friends group will automatically be tagged with the Friends tag when new images are imported. Although this would be better suited if Beagle and FSpot shared the same database.
This sort of thing would probably benefit from a storage mechanism like WinFS where you could associate data with other data but there is no reason that this couldn’t be provided by a separate database.