NCED and other projects encode a lot of metadata in the pathname/filename. Seems like a general extractor that can pull info out and create structured metadata would be very valuable in the 21rst data set, data consultant, and rich collections use cases