https://isnot.at is a lexicon & demonstration site for a proposal I made a while ago for micro-reviews.
It lets you review any subject represented by an at-uri with any number of (direction + adjective) tags, eg. “The Hitchhiker’s Guide to the Galaxy is really funny”, or “Puzzmo is cool”.
I thought it might be a good way for Atmospheric apps to model tagging/thumbs/stars/reviews so they can work across apps, and allow both simplicity and nuance where it’s needed.
The next things I’d build would be data-science based; xRPC endpoints that recommend adjectives which would help you explore your opinions on a thing (Marmite: is/not delicious?) and, if this takes off, recommendation algorithms based on the aggregate dataset. (it even support multiple locales, and could use k-means clustering to make the actual locale of a tag irrelevant, it’d just look at what tags you use for the things you like, and match that to the aggregate).
The docs show how it works, with an interactive demo. Let me know what you think!