The FEP process: rough consensus and running code

[strypey]

Continuing the discussion from FEP-be68 Audio objects:

@Funkwhale

After discussion with @mirlo devs we are dropping the concept of channels in favor of a more simple approach : associating an artist object with an actor account. The FEP has been updated.

FWIW this is why I think the FEP process is so valuable, and why I keep banging on about it. A few weeks of trying things out in code branches, and talking to each other about how it’s working (or not, as the case may be), seems to be an effective way of finding protocol consensus.

This is why greyhair hackers talk about ‘rough consensus and running code’. Because we’ve learned from bitter experience that abstract discussions, prior to actually trying things out, can go on for months or years with no resolution. They can become very circular, and often heated as a result. I think this is what the infamous blog post about “architecture astronauts” was handwaving at (somewhat ineffectively IMHO).