Can you clarify again what is this fixing? I went back into the old discussions but it's hard to find the actual problem.

@vitorpamplona Relay hints. Putting them on the d tag breaks conventions. People already have code that does things with "p" and "e" tags automatically, since they're considered "standardized tags". Querying with nak for example has special CLI flags for #p queries. Another example is using p-tags to connect nodes in a graph database. Semantically, it just should have them. This method ties them together. But it's only useful if people commit to doing it strictly so it can always be depended upon.

ohhh interesting..

Ack

In pretty much every case I've seen it's better for d tags to be random. Maybe p/e tags would make sense though, since they're immutable, but what's the use case?

@staab In Ditto I need to manage records about other users and events, about things that don't make sense to use labels.

#1534 is a good example. Also: https://docs.soapbox.pub/ditto/nsec#user-records

This probably doesn't deserve to be a NIP on its own, but #1510 exists so I went for it.

I think #1540 could be a better alternative