Hi,
No problem, Google's failure to tell me you replied delayed my getting back to you to. :)
We've added a branch in the bookmarklet to import videos and scrape metadata from
mirc.sc.edu . It works great, but all the metadata we're pulling from the source page ends up in a json blob attached to the video in MediaThread. In my ideal world, filter fields would be shown in the interface dynamically based on the metadata that exists in any given collection (or subset of a collection). So, since MIRC has fields for things like Geographical Coverage, as long as they show up in that json blob on import, any collection with a MIRC video in it would have a filter field for Geographical Coverage.
I know this is potentially problematic from a database perspective, but my thought would be to treat any of these metadata fields very naively - as long as the bookmarklet passes a field with some data in it, MediaThread trusts that it is what it says it is. No need to try to normalize between MIRC's Geographical Coverage field and someone else's Latitude and Longitude fields.
Somewhat lower hanging fruit might be to at least designate that a special named field in the JS blob (like, er, "tags") is automatically parsed into MediaThread tags rather than just left in the json blob.
- John