[META] IPNS/Namesys Todo

You may have noticed that IPNS, while very handy, has a number of ongoing issues. Let's use this issue to track them. Before adding in new bugs I'd recommend reading through the existing list.

We'll likely need an overhaul of the IPNS/Namesys APIs

- [x] IPNS over PubSub still requires the DHT for resolving non-self/identity keys #6360
- [ ] IPNS TTL/caching Issues #7564
- [ ] IPNS over PubSub caches the latest record for each key indefinitely
- [x] IPNS Republisher ignores EOL #7537
- [ ] Resolved IPNS/Namesys records are not cached/stored in the datastore (i.e. does not work when restarting in offline mode) https://github.com/ipfs/go-ipfs/issues/8245
- [ ] IPNS pinning (e.g. #4435)
- [ ] Pin IPNS keys so that we search for the latest records upon restart
- [ ] Recursively pin/unpin the data pointed to by an IPNS pin
- [ ] IPNS records should be republishable by third parties without requiring them to have the private key #3117
- [ ] Come up with a better way of describing the distinction between Namesys, DNSLink, IPNS and `ipns://`
- [ ] Allow `ipfs name publish` to return after either IPNS over PubSub OR IPNS over DHT publishes are successful, instead of waiting for both
- [ ] Have IPNS over PubSub's PutValue function wait until at least one record has been put on the network

----

Note: There are other issues, irrelevant to this meta issue, effecting IPNS publish/resolution performance including
- Stabilize PubSub and IPNS over PubSub so they are no longer experimental
- Various DHT/libp2p related issues including
- Getting the nodes acting as servers on the public DHT to be running the latest versions of go-ipfs
- Could be helped with https://github.com/libp2p/go-libp2p-kad-dht/issues/616
- >= go-ipfs v0.5.0 servers will not choose to act as servers if they are unreachable which is a big plus
- Mandating that DHT servers support QUIC (will decrease DHT lookup times)
- Adding QUIC support for session resumption (will decrease DHT lookup times)