To me, unified search, across any audio hosting/ social music services artists or audiences choose to use, is the most obvious benefit of a federated approach. So glad to see this being enthusiastically explored here.
From an audience POV, I know I’m always drawn back to the bigger search portals (BandCamp, YouTub) because I’m most likely to find something like what I’m looking for. Diversity of hosting choices has many upsides. But searching dozens of niche sites one by one, and finding mostly tumbleweeds, isn’t a sticky experience. It’s not that much fun for DJs and audiences, and doesn’t do much to help artists get discovered by potential future fans.
So let’s say there’s a federation of audio-hosting and social music services. Each with a search tool that can search across the whole network. But what happens if people search for an artist that isn’t hosted within the network? Rather than tumbleweeds, what kind of results could they get that would respect consent (and thereby copyright), but still be useful?
Let’s say the federated search space included a complete index of all released music, with links to all the places I can hear it, and buy it. Existing release index projects (eg MusicBrainz, Discogs), and music information projects (Wikipedia pages, AudioCulture.co.nz) could be integrated into the search space, if their custodians are wiling, or new ones could be set up for the purpose, with their index under a pro-sharing open data license.
That way, if I go to a FunkWhale or BandWagon service, or to Mirlo or whever, and search for an artist or album hosted within the federation, I could get the detailed profile curated by the artist, with the music right there to listen to. But if I search for Rage Against the Machine, instead of getting nothing, I could at least get a copy of some publicly available information about them, and links to some places I can listen to or buy their music.
Maybe with a set of ‘if you like … you might also like’ links to artists hosted within the network, chosen by genre tags, newest to oldest (or vice-versa, or randomised). Or if you want to get really experimental, by an algorithm that looks at what people say they’re listening to across the federation, sees that people who listen to RATM also listen to artists A, B and C, and lists them under ‘you might like …’. Or the person could toggle between these 2 serendipity modes, and maybe others.
The downside of this, for anti-corporate radicals like me, is that sometimes patrons and their money would leak out, back to the corporate platforms. But I think this is more than counterbalanced by the fact that searching on services in the federation would always be useful to audiences. So we’re much more likely to keep coming back, to a place where independent music results present less friction than corporate ones.