-
@ David
2025-06-16 13:33:42
i've been thinking about uses for it also, it occurs to me that probably i could create an index for the nilsimsa hash of the content field of nostr posts that have text in them (notes, wikis, long form, etc) and then you could make a search function that lets you feed in an arbitrary length text string input and it could return all of the events whose nilsimsa hash was within a specified distance.
this would be a pretty heavy compute search function i guess. but then also maybe a less intensive form would be to analyse one npub at a time and generate the nilsimsa hashes of each and perform a matrix comparison of them all to derive some number that places them somewhere in the finite field of 256 bits.
probably could just add them together using a modular multiplication operation like used in signature functions to derive a coordinate that could be used to find the hamming distance between different npub's text and form an unbounded (spherical) 2d graph that becomes a map you can traverse to find people with similar content as each other.
you could also then analyse another single-author attributed collection of text from somewhere anywhere else and also place them in the map.
not exactly sure what this could be used for but probably it could add a lot of colour to a web of trust graph too. and maybe help people find other npubs that talk about similar things or similarly talk. so to speak.