Generic design for automatically updating metadata from file?

I've created a package to provide (automatic) updates of metadata (and thumbnail) from different document types (so far MS Office & PDF docs) to Plone(4) content. It comes with an example content type, and normal (plone.app.contenttypes- based) File content type can be used as well (need to assign two behaviors to it first). There's a custom view as well.

Document-specific support for metadata extraction is provided by named adapters (where names are simply mime types) adapting plone.namedfile INamedFile to a metadata interface. An event subscriber then copies over the metadata if an adapter is found. The actual document type -specific extractors (adapters) are provided by other packages (Products.OpenXml & collective.pdfdocument).

I guess the plone.namedfile NamedFile field could have been extended instead, and maybe that would be better?

Also, there are some similar efforts, such as unweb.iptc for EXIF/IPTC (image metadata).

Rather than file type specific implementations proliferating, it seems to me it would make sense to provide a generic pluggable mechanism for file metadata extraction & updating. Such as what I've implemented. But given how easy it is to roll your own (just a subscriber is needed at minimum), I wonder if anyone would use such a generic packgage? Even if a standard way to access metadata would be nice.

Or, maybe a PLIP for common file metadata API should be created?

Opinions and comments welcome. Thanks for reading this far.