Decision on peer IDs, CIDs, multibase, and ed25519

[Stebalien]

So, the issue of peer IDs came to a head today when I learned that OpenBazaar can't update to the latest libp2p because we've started auto-inlining ed25519 keys into peer IDs (changing the peer ID format for anyone using ed25519 keys).

We also need to support:

• Arbitrary hash functions. This means we really need some field in the key to specify the hash function.
• Multibase. Browsers need this for IPNS.

I have three concrete proposals on how to attack this but we need to make a decision:

• multibase-peer-IDs: Specify peer ID hash function in public key specs#111 (comment)
• CID-peer-IDs (text only): Specify peer ID hash function in public key specs#111 (comment)
• CID-peer-IDs (everywhere): Specify peer ID hash function in public key specs#111 (comment)

This hasn't gotten anywhere after years (?) of grumbling on GitHub so I'd like to schedule a synchronous call to make a final decision: are we going to use CIDs as peer IDs or not.

In addition to the libp2p team, @whyrusleeping and @daviddias really need to show up.

[daviddias]

CID-peer-IDs FTW