AI should not be tied to the cloud. Privacy makes things more convenient instead of difficult as most big tech likes to tell us
I will put this under "relay discovery" feature set.
Good question.
1. Blowater's relay list can be modified and sync across device.
2. It uses a CRDT approach which you watched last week. It is not compatible with NIP-65 and is more stable than NIP-65, meaning other client can't mess it up. I will draft a NIP to let more client adapt it.
3. Yes, Blowater should somehow discover a npub's relays in a "smart" way, be it NIP-65 or other methods. Then the user can choose to copy these discovered relays to Bw's relay list.
On Android, Firebox and Chrome "add to home screen" supported.
On iOS, Safari "add to home screen" supported
Go to https://blowater.app and play with it. I will zap anybody 2100 if the issues is not duplicated from current GitHub issues.
I have 42K to zap. Valid until Sep 15th.
Sorry for the delay, https://github.com/BlowaterNostr/blowater/issues/47 has been merged!
It's 2 months late but a lot has happened in the mean time.
Blowater team's producitivy is back online!
@me if https://blowater.app has any problems. We just released a big refactor
Indeed we need to improve the onboarding. A redesign has been planned. https://github.com/BlowaterNostr/blowater/issues/127
I accidentally introduced some bugs several weeks ago and have fixed them now
test
Utility software design can learn so much from game designs.
Nostr is good, so join our X?
Always back up your keys.
Yes. O(n) in memory is faster than nlogn in disk for small data such as 100K events. That said how horrible IndexedDB is. Loading the whole table in indexed db is also faster than select 1 by index. My guess is that V8 team and other chrome component teams are at different levels. V8 is so fast that most computation heavy thing can happen in JS now.
Compatible?
https://github.com/BlowaterNostr/blowater/pull/160
After almost half a year of experiment in production, we decided to sunset the "Chunked Image Events" feature in Blowater. This feature allowed users to send large images over relays in DM in an encrypted manner.
This feature predated NIP-95 and was probably the first in the community to experimentw with relay stored encrypted content.
We sunset it for 2 reasons:
1. It broke the atomicity of events. An image or a file should be an atom. Breaking them into many events broke this pattern and make the implementation difficult to maintain. An event should represent a whole thing, instead of a chunk of a whole thing.
2. Performance is not good. But performance is actually not as bad. The real problem is that sometimes relays lost events so that some chunks of an old image might be lost and thus the client will never be able to recover the image.
But, encrypted binary content in DM is not going away!
We will implement a standalone binary storage service for DM, similar to nostr.build. Just like Blowater itself, we will first experiment with it close sourced. Once we feel it's good enough, we will open source it.
My vision is to implement several side-network that deal with specialized tasks that can interop with relays either directly or through clients.
This binary storage service might also just be decentralized as well, but it's not another IPFS. We have different and more realistic goals on Nostr.
Stay tuned.
