no, people need to not overweight NIPs or the NIPS repo so much; you can implement whatever you want on nostr without every touching the NIPs repo.

My process when I want to come up with a new use case is:
* search the repo and nostr for conversations about my use-case.
* look for a kind that isn't used (REQing random numbers I like)
* make up the schema I want.
* start using it.

* After a while I might create a NIP to document it so other people know it's there.

Both first and last step, which might touch the NIPs repo are optional. Particularly if a use-case takes off, the last step is largely irrelevant since the real-world implementation is the spec.