nostr:nprofile1qythwumn8ghj7anfw3hhytnwdaehgu339e3k7mf0qyg8wumn8ghj7mn0wd68ytnddakj7qg3waehxw309ahx7um5wgh8w6twv5hsqgzxpsj7dqha57pjk5k37gkn6g4nzakewtmqmnwryyhd3jfwlpgxtsnyw5n2 nostr:nprofile1qy2hwumn8ghj7un9d3shjtnyv9kh2uewd9hj7qgwwaehxw309ahx7uewd3hkctcpz9mhxue69uhkummnw3ezumrpdejz7qghwaehxw309aex2mrp0yhxummnw3ezucnpdejz7qg3waehxw309ahx7um5wgh8w6twv5hsqgpjuxp8vd29p6ancknaztql3eajk52y8xkppfn7au7elkw9c68zg5w6awvc nostr:nprofile1qyghwumn8ghj7mn0wd68ytnvv9hxgtcpz9mhxue69uhkummnw3ezuamfdejj7qgmwaehxw309a6xsetxdaex2um59ehx7um5wgcjucm0d5hsz9nhwden5te0wfjkccte9ekkcettw5hxgetk9uqzphtxf40yq9jr82xdd8cqtts5szqyx5tcndvaukhsvfmduetr85ce4znfpm your considerations would be appreciated.
Discussion
We've been building for over a year, with a whole set of new kinds, without a single NIP for any of them, and nobody has yet died.
I find the kind list and tag list on the readme most interesting, to make sure I don't accidentally reuse a popular number, and to peruse existing data structures. I sometimes dig stuff up out of PRs, like public messages, and implement it. I don't bother to wait for a NIP.
Either it's a good idea, or it isn't. The bad ideas eventually just peter out, and I decide for myself which ones are the good ones.
New devs often ask me for advice on NIPs to use and I direct them to open PRs that I'm following, but they're sometimes scared to use them.
NIPs are fine. Most people take them way too seriously though. It's just a place to interoperate. We need a place for that.
To me, NIPs are a reflection of usage. If that is what we want to do, we can create a portal where each client/relay declares their support for NIPs and the most supported one become an official nostr list. If clients drop support over time, they disappear from the list. A backwards way of developing specs, but one that trully represents what Nostr is about.
I agree, that's why I want to get rid of them.
The way they exist now gives too many wrong impressions. People see a central registry of NIPs and think that is "the Nostr protocol", when in fact many of those NIPs are not used already (or most people agree they shouldn't be used) and we can't touch anything on the list because it is a very serious list. At the same time there are many kinds being actively used with multiple implementations that aren't merged on the repo, and, most importantly, the way in which each kind is or should be used is not in the repo either.
Once the "official NIPs" aren't a thing anymore someone (or actually many people) can create independent registries that track implementations or do the job the current NIPs repo is doing, but much better and in a more fluid, opinionated way.
I don't know... I was hoping we could have a better portal for NIPs inside of Nostr. But all implementations for that kinda suck.
Yes. Probably because NIPs are the wrong abstraction. But also because we're stretched too thin trying to make all the things "on Nostr" at the same time.