Oh, we add "type" for Bible, book, magazine, academic, documentation, blog post, poem, etc. to determine how to render it, prettily. Using that or "t" tags like "education", or "development" to find books about specific kinds of data or from specific fields. So, that would be interesting.
We have the "author" tag, for adding the plaintext real name, like "Aristotle", so, that's always interesting.
Also the language tag and a fuzzy search of the title tag would be useful. I have to do too much fuzzy filtering client-side and it's slow.
Another thing would be to be able to say "from the @harvard.edu domain" or "from the @gitcitadel.com" or whatever NIP-05 domain. So that users could narrow to particular source.
Also, m and M tags. I'm adding those, but some human-readable query is nice. "Should be an index" or "should be in Asciidoc format".
That sort of thing. There are a gazillion tags in the publications, but no real way to get to them, at the moment. I just pull hundreds or thousands and plow through them, with a client filter.
