One issue I see is that NIPs don’t go through stages like TC39 proposals do. NIPs require at least one working implementation, not that most major clients implement it. It would be counterproductive to require major client or relay to support every single NIP before it becomes official.
Discussion
* requires two implementations
One implementation is not enough.
It needs at least 5 recognizable implementations with all 5 clients commenting on the GitHub NIP signaling full implementation.
1-3 clients is not enough.
No merge process is too loose.
What counts as an "implementation"? It's all subjective at the end of the day. Bureaucratic processes just exist to create confusion and extra work to cover the actual process, which is not and cannot be defined and changes as the power dynamics change.
Agreed.
If i make a PR to add a NIP about some funny Kind i invented. Then me and some other developer decided to implement it in our apps...does that consitute as nostr now and gets added to the list of NIPs.
There must be some decision making process about adding NIPs. whatever that process is, can it be stricter?
Exactly.
Why do you want it to be stricter? What NIPs have been merged that you feel shouldn't have?