How about a decentralized list of musicians, organized by genre, where the list of genres is also a decentralized list?
Anyone can add an item to either list, and we use WoT to make sure spammers don’t screw it all up.
Discussion
HPM — tell us more about the WoT music scorer you mention. What do you want to see scored - songs? Artists? Will this be for users to find new music?
Yes, I believe a primary use case would be new music discovery. It boils down to, if my music profile matches closely to that of another user, and they interacted with 9/10 same songs as me, it's a good bet ill like that 1 other song on the list. Even stronger if that user rates it highly. Now extrapolate this across a network of interconnected users.
There's a couple ways this can be implemented.
The key in any scenario is to generate individual profiles by npub. You could then use the tagging system for genres/styles, artists, songs, etc. For each one, points are scored for positive interactions (boosts, zaps, likes, etc) and thus builds individual scores, per user, per category.
If this system is build by a third party service, any app can API into this and utilize the same scoring engine. Multiple engines would decentralized the process as each could have a different weighting model. Apps could allow users to select different models.
Visually I think of it like those stat charts in an rpg like this:
With each variable being some genre or tag or artist or whatever. Basically create a matrix variable and that serves as an npubs unique music wot score. A bot can log activity similar to how nostr:nprofile1qqs0q5wwke7qr4j7d4sk37dmf4mzytzy2r59w47f8yc6627rzm9urjcpz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0dsq3samnwvaz7tmhv4kxxmmdv5hxummnw3ezuamfdejsstjm9w bot listens for boosts.
The closer the score the closer the style preference. You then extrapolate suggestions based off similar tracks that haven't been interacted with.
I see your point with all of this but you're trying to interpret data that doesn't exist.
We need ways for people to listen to this music before we can worry about seeing what your friends are listening to.
That said my app stablekraft.app does have some Nostr stuff added to it. Status updates for what your listening to, favorites (bookmarks) that are stored on Nostr and I use the metadata proposal that nostr:npub1unmftuzmkpdjxyj4en8r63cm34uuvjn9hnxqz3nz6fls7l5jzzfqtvd0j2 uses with Fountain when a user sends a boost and those even show up in Fountain b/c its the same spec.
https://primal.net/merryoscar/open-podcast-payments-with-bitcoin-and-nostr
Podcasting 2.0 uses GUIDs to identify items and thats how Fountain knows what song or album I'm Boosting.
We now have a couple decent options available. I agree we still need more ways to listen but a huge part of retaining users (its not enough just to attract first timers) are vital ancillary features like new music suggestion. Tbh i think thats essential in this age of music if you want to attract and retain most modern users.
It's not decentralized but anyone can use it (though it's interested)