Not that it helps much, but a few comments.
Right. To be able to extract meta-data from images would give the edge towards full-pro lightboxes. This is what the “big photo apps” do.
Use case: 4,000 paintings of Bonfire Night in the UK.
FYI, in that use case there is NO way I’d ever need or want TW to permanently import the meta-data. Merely dynamically read it for a lightbox.
Here all that is needed is TW selects a number of images to present and temporarily retrieves meta-data from, say, 50 images, for a lightbox.
The point is the images already hold the data so why add it again (for that use case)?
In this case I already conceive the “image” as a kind of “external tiddler”
Just a comment
TT