do you prefer sending binary over the WS as opposed to the other proposal[0] that handles uploads with a multipart-form? if so, why?
Discussion
I believe a WebSocket connection will be superior to any multi-part HTTP design in the long term.
That doesn't mean NIP-96 is bad or can't be used right now.
I am a full supporter of getting them all (NIP-95, NIP-96, NIP-97) merged and letting the best design win.
We just must not keep the current state of image hosting services in Nostr. That will be a disaster in the long run.
interesting, I had not considered a future where the protocol supports multiple ways to do the same task. is there precedent for that in nostr? you know the nips better than me but to date it seems like there’s one way to do things, generally speaking.
It's not that rare. NIP-26 and NIP-46 are kinda doing the same thing by very different means. Kind 3 relay lists and NIP-65 lists are "similar". Kind 10000 for mute lists is also done as Kind 30000 with a `mute` label. NIP-04 and the new NIP-24 will both do Direct Messages.
As Nostr grows, we will see more and more alternative ways to do similar things.
🫡
something needs to happen to decentralize and stabilize image hosting. our current file hosts are doing an amazing job handling this task, but it's really confusing to say that our text is censorship resistant, but our images and videos are not. i don't know what the best solution is, but we do need something long term standing.
hypothetically- the person(s) who care about the image should host the image. that aligns incentives. not every image needs to live forever and only individuals can choose which they care to retain. the issue is today it’s not easy to self host content. but yet I’m typing this from a mobile phone with over 100GB of storage and an internet connection.
I'm hoping with storage costs continuing to fall, personal servers like Start9 and Umbrel will become more common.