@silverpill

> Could you add a new category that represents clearnet instances connected to Tor federation?

This is actually difficult to detect! Automatically dedup/merge is going to either be slow and result in some false positives or reasonably fast and result in a more false positives. It's not intractable, but getting it right (and making sure it doesn't go off the rails) would be non-trivial.

All the data in the system is a moving target, so it'd have to be fuzzed a little. Maybe I could do something like generate a fingerprint based on a hash of the instance description plus (truncated to account for wobble) ratios of users/posts, posts/peers. So then you can flag pairs by matching the fingerprint, and then compare the numbers, see if they're all within a certain range of each other, then mark as a duplicate. It's something that has to be done carefully to be done correctly.