Community apps feel like the new nostr gold rush, everyone keeps building new ones that all work completely differently. I agree with their importance (obviously), but it seems like this has the potential to fragment things more than ever, just because communities are a rorschach test. Everyone also has their own libraries and preferred development stack. What if we just all agreed to start from scratch and work together to create something new? It probably wouldn't work because we're all so disagreeable, but one really good community app would be better than 10 incompatible ones.
Discussion
I genuinely don't understand this, but are they all operating from the same NIP? Whatever one is for "Reddit-like moderated communities"?
I agree, this is something that should work better. I just see it as an issue of user base. As small as Nostr is, ~15k daily users, that's just for the "Twitter-like" part of Nostr. Take another 1 or 2 percent for the "Reddit-like" part and it gets even smaller.
But I guess that's all the more argument to get the technical architecture right before crowds show up...
Yeah, communities can be an on-ramp for people. But the apps have to be good first. NIP 72 has the most implementations, but the spec is kind of garbage from a moderation stand point. It's fixable, but also only applicable to reddit-type communities (and has all the same problems with tyrannical moderators).
You mentioned nostr could have been built using a binary format and CRDTs in your book. Have you thought about how CRDTs could have been implemented instead of events for things like follow lists or was it just an example?
A little bit, I'm not super proficient with CRDTs, but it's not too complex to make a Set data type. Right now follow lists are completely replaced every time, making them quite brittle. A set with add/remove operations would be very simple, with the cost of taking up way more space. I proposed this over two years ago: https://github.com/nostr-protocol/nips/pull/349
The nice thing about this is you could build follower lists more cheaply, but it would be much more expensive to build follow lists, since you'd have to download everything and reconcile it in-memory. Snapshot events could help with this, but nostr's event model is just too simple, and relays too dumb to solve this very well.
1 Year From Now: nostr:npub1jlrs53pkdfjnts29kveljul2sm0actt6n8dxrrzqcersttvcuv3qdjynqn has received the first OpenSats Pubky Grant!!!
Sometimes you need a company of sorts to herd the sheep
Isn’t this exactly what we all ought to want? I could care less that a bunch of peruvians are chatting with a bunch of Chinese via Damus while I’m over here being a dork on Primal. Same with Bitchat Mesh. Get weird, man!
Well no, because sooner or later all these things will need to pay for themselves. Devs will need to earn their own incomes. Infra costs will need to be met.
Scattershot is fun but incredibly cost-inefficient.
Hard disagree, sorry. I think that we absolutely need the most amount of diverse apps doing whatever they think is best, otherwise we will just recreate the centralized internet but this time using nostr.
are you retarded?
his point was that different apps should interoperate
lol I know that, and I'm saying that it's not necessarily needed, we can build DIFFERENT things, even ones that don't interoperate, and that's still fine
what about you ignore the trash and just work on the good stuff?
the stupid ones won't last