Nostr Web Client

Found a way to do e2ee multiwriter key value store, that can be searched and replicated just as efficiently as a non-encrypted one, and synced as efficiently as an append-only log, but allows deletes and garbage collection.

Single rountrip to full sync

Log n roundtrips for partial sync

Log n proof overhead (Nostr is linear)

Authenticated ranges

Really really good.

Combined with a more pragmatic discovery and client-sever architecture, it will stay be order of magnitude more complex than Nostr, but order of magnitude simpler than p2p, so actually manageable and adds tons of features that makes it worth it.

If sovereign cloud hosting is going to be viable, we need sub-linear scaling and we will need e2ee (so relays don't worry about content they host).

None of this is going to be ready anytime soon, but it I am convinced it is worth it.

It is going to be hard to maintain the motivation and getting people to care though 😕

Nuh 2y ago

I am 100% sure if I did Fiatjaf will fight it because it is complex, like he did with Nip26

I get your point, but just ignoring all the technical limitations (for example relays don't understand what are orphaned nodes to GC it), and even ignoring the overhead of linear size of signatures.

Ignoring all that, whatever I build using nostr events will be opaque and useless to Nostr unless they use my very code, so the excitement won't translate, rather the values of the community will reject the complexity and fight it.

The remaining upside of Nostr is basically free relays with no friction to sign in (albeit unreliable storage), but the moment someone else offer better cloud storage, you are left only with the tech debt of the choices Nostr did.

Case in point, Nosdav has ZERO reasons to support Nostr event shape, or even secp since it doesn't really run in Nostr infra, you are just trying to get Nostr devs interested this way, but I am not sure it is a bet that will pay off.

SimpleX did its thing, and still got biggest Nostr supporters attention.

Maybe you are right, and maybe there is not much money or financial gain in this space anyways, so might as well do what you want to use yourself, and don't worry if it will get devs attention.

Reply to this note

Please Login to reply.

Discussion

Melvin Carvalho 2y ago

Yes, I built nosdav because I want to use it myself

Remember I built Solid before that. The issue with solid is that the pub/sub over websockets had a bug. So my realtime chat app was not working across domains. Even though I fixed the bug for several years, it did not get prioritized upstream.

So I decided to use nostr instead. And we built nostr up to be bigger than Solid. To give you an idea, it took me one full year working 24/7 to onboard 1-2 developers to solid. In nostr there are 100s. Because it is simple. Solid started simple, but academics came in and made it more complex. Now its hard to get any non-academic developers there.

Nosdav is something I want to use. So I'm simply porting all the best parts from solid to nostr. I can only do it one peice at a time as solid was many years of work from dozens of people.

I like taproot and schnorr (in fact solid is implementing them now) because it makes it easier to integrate with a world class open source payment system. I always wanted that for solid, I even started a project called solidpay but it didnt catch on. Addionally we tried lorea before to be social + payments with pablo and amir. Amir then came to bitcoin and made the BIP system. So it's all been one long journey to get to nostr.

Now we have nostr, it's brilliant. I love it. It's everything I wanted and knew was possible. Porting the things I like such as personal storage and micro apps isnt going to be that hard. And it's going to be good enough for 99.99% of users and have great utility. It'll be better than apple and google at least. There's also tons of use cases that fit the nostr eco system. Because relays were not designed for long term storage, they were designed to relay messages from one user to another.

Nuh 2y ago

I am very well aware of the appeal, but:

1- I don't have to make Nostr architecture my base just to use free pub/sub. I can just take advantage of it when I need, especially that friction and costs are zero.

2- if all I need is 1-1 relay, then there is also the competition from SimleX, which is definitely more private, so Nostr now is best only for many to many broadcasting.

3- there is only so much I can do with ephemeral data, and I want persistent and more importantly e2ee, nostr is not made for that, and that can't be done without increasing Nostr complexity, ok but whatever data store will hold my data persistently, it can also double as ephemeral open pub sub if ever needed. Especially with pkarr when I KNOW where my contacts store their data and can send them messages there, or fetch their latest public notes.

Nostr relays may very well end up being the RSS of this ecosystem, the aggregators and feed generators that people subscribe to to get a precompiled view, which I feel Fiatjaf is trying to cement as the way to use Nostr and I full agree with that, but it can't be the infrastructure for what I want, not without ruinng its simplicity, case in point, you could have built NosDav over relays, but it would be a bad idea.