Has anyone looked into pear much? Curious about opinions. I'm skimming the docs and getting the same vibe I got from previous P2P things. It's all either made up or I'm too dumb to understand it. I would like to see some kind of network topology diagram or explanation.

Reply to this note

Please Login to reply.

Discussion

Also getting Urbit vibes — so many micro-libraries: hyperbee, corestore, protomux, hypertele, on and on

lol. Let’s start doing Kelvin releases.

sounds like the components of libp2p also

I stopped trusting keet when I saw there and FBI cooperation

I'm glad I never used Keet. I was waiting until they open sourced it but then I read their privacy policy and decided I wouldent touch it with a 40ft pole.

Yeah I’m definitely pass

What's wrong with the privacy policy?

Sorry I used the wrong phrase its their tos. tldr despite being built on open tech (hole punch) they will sue you if you reverse engineer it and try to figure out how it works. They say they dont collect data but with a "balck box" we dont know for sure and I could be misremembering but I thought they talked about open sourcing keet. So I wanted to wait to use it until then but after reading the tos I doubted that it was likely and decided to stay away from it.

But pear is open now & that is what uses holepunch & hypercore. Keet apk is coming with the next major release.

I really dont care about keet. Having an apk wont change that its not open source. If pear is actualy open source then cool. 👍

What FBI cooperation is there?

Well, i guess you can say whatever you want if they aren't open source. The burden of proof is on keet.

Pear runtime is now open, Keet apk is coming, supposedly with the next major update. It's pretty clear from the way it operates that it's truly p2p tho. If I'm not online things don't propagate. I have to leave an instance running at home to seed my rooms. And everything tends to work better when there are more peers. But blind forwarding is also on the way, so media should propagate more reliably when that hits.

btcstoring.com is my domain name and is for sale at 0.50 BTC

I am jacks complete lack of surprise

So what would you do if your centralized service was targeted by the govt?

Personally, I'd start building ways to decentralize everything else so that targeting me would ultimately do them no good 🤷‍♂

For me it’s just a 🚩 not to trust it

Code is open, no need to trust.

Sorry but I don’t use spooks products

sounds like 🚩

Listening to your podcast with fitajaf made me think about how pear compares to nostr. I am sure they are different things

But… nostr:npub1h8nk2346qezka5cpm8jjh3yl5j88pf4ly2ptu7s6uu55wcfqy0wq36rpev did a podcast with the pear team few weeks ago and they were making a lot of comparisons of nostr and pear… I know Guy uses Keet a lot

Is it true that nostr and pear are a new decentralized internet attempt?

I'd say it's a continuation of the same project that's been going for years. Nostr built on scuttlebutt, pear is building on DHTs and other related technologies.

Very much so. Maf was working on BitTorrent and basically moved straight into this architecture from there, it just required tons of modular units because it is a full replacement for like 3 or 4 other major protocols of the internet stack.

Woah, I’m learning a lot here, thanks?

I love typos… *thanks!

Oh, I need to learn more about all this, super cool

Not ‘new’.

‘Another’.

you can disagree, but 1st second vibe tells the truth a lot of the time.

I was looking into pear aka hypercore quite a bit in 2022, a lot more than nostr at the time before getting pulled back here.

Pear was previously named Hypercore Protocol and has been around as a research project for a while, with a relationship to Bittorrent.

It uses the Kademila DHT to discover peers on the internet but requires holepunching.

It is worth looking at the internet archive to see the old diagrams before they modified the website to be what it is now.

https://web.archive.org/web/20220704132854/https://hypercore-protocol.org/protocol/#hyperswarm

As for quality of audio/video over keet, its much better than any other app, and also works over VPN.

A lot of hiccups last year caused keet to break on my network but it seems to have been fixed in the last few months.

The quality of experience still needs work, but I'd not dismiss it entirely and keep an eye on what is happening. There are enthusiasts using it on a regular basis.

It was called Dat Protocol once, then they rebranded. Paul Frazee tried building a decentralized social network called CTZN using it but don't think that took off. Paul has also heavily contributed to the SSB ecosystem, and now works at Bluesky if I am not wrong

https://www.youtube.com/watch?v=wL44nU_eAhM

The better thing to do would be to use a simpler protocol for private sharing of files etc (like the Signal protocol) and announce your preferred servers using nostr. Kinda like how live streams, audio spaces, and git stuff work over nostr

The same Dat that Max Ogden started? I always liked that guy.

Yes. That same Dat Protocol was renamed to Hypercore and was governed by the Hypercore Protocol Org. Now it's governed by Holepunch, a company that's "backed by" Tether, so it's a subsidiary I guess, just like John Carvalho's Synonym.

Fun fact: John Carvalho's comments on some podcast that they will do a decentralized Twitter alternative some day (probably was thinking of using Hypercore back then) made me realize it's possible, and then I discovered nostr.

Paul is now the lead dev / architect on Bluesky.

he also did beaker browser, a p2p web browser built around hypercore/dat.

Yeah I’m a big fan of Paul. It’s nice to see him getting success with his work on Bluesky.

Awe. 🫢 thanks 🙏

Thanks for connecting all these dots. I hadn’t realized the history of Pear. I did just find this interesting post-mortem from Paul - sounds like the project failures were more around design than around the protocol itself.

https://paulfrazee.medium.com/building-in-the-open-70ac9dccf1aa

The one that annoys me about Pear is that its primarily developed in javascript and nothing else afaik. Development is super slow, not like nostr where its a new NIP every day.

For some stuff Nostr works, and for others it doesn't. we need a better implementation for DMs and any other type of p2p.

Nostr is not going to solve all of our problems and if we have to cherry pick from different protocols to serve the purpose I don't see anything wrong with that.

Anyway, feel free to debate or weigh in, anyone else who has spent time reading the pear docs.

Doesn’t nostr:npub1ajv7m32k0cpgzha32qszsh304qusjvwwmavus0ttktzldms4xzusuftppj have a group chat or something on pear?

I have multiple keet groups. The current lack of blind forwarding is the only real complaint, & that won't be a problem much longer. Because of that I have to make sure the pics I upload from my phone propagate to the laptop I leave running all the time as a seeder for my rooms so that others can see them. The whole thing is really amazing tho. My brother is building his own tool that runs on pear. Hypercore & holepunch & pear runtime are going to completely change the internet IMO.

I think the whole hypercore / pear / dat ecosystem is both super interesting and frustrating. The JS only part is a major problem for me, and it is pretty complicated.

It’s an alternate universe to what protocol labs and ipfs have created, but with less funding and no token.

It’s literally built like BitTorrent and it’s over 10 years in the making so it has tons of moving parts. But Pear Runtime is open source and is much easier to build on and obscures away a lot of the hyperbee, hyperswarm, etc stuff.

All the naming stuff is confusing as hell, but it’s slowly coming together in my mind. I think there was just a ton to build to get this to work and they built it in a very modular fashion and their naming convention opted for unique instead of intuitive.

What I can say is it does work. It is P2P. And I think it’s one of the most

Important things being built right now alongside #Bitcoin, #Nostr, and #Ai.

I strongly suggest not dismissing Pear.

Do you have any opinions on how hybrid nostr/pear apps might work? DHTs are ok for indexing, but can't hold all the data. Maybe pear could be used to replicate indexes of the nostr network (i.e. kind 10002 and maybe kind 3's, what purplepag.es currently does). Nostr keypairs could also maybe be used along with that index to bootstrap p2p DM channels. But does pear work offline? IOW, to receive a DM both peers would have to be active I assume. So nostr's relay idea could be used to bridge the temporal gap between peer availability.

this is part of what nostr:npub1t89vhkp66hz54kga4n635jwqdc977uc2crnuyddx7maznwfrpupqwra5h9 's merkle dags could help with, creating high level, divisible units of the branches of related trees

at this point the outbox model (like nostrudel uses) is the best solution - advertise to the network where to contact you and where you send your messages

The data layer is purely agnostic, meaning everything about #Nostr can work exactly as it is now, but be communicated and connections made via the DHT. (The DHT is solely for peer discovery, same as torrent networks)

The Nostr protocol is actually completely agnostic to the relay system. Relays is just the low hanging, easy method for distributing the data layer. It could run on Pear just as easily (conceptually, obviously implementation would have some hard work behind it), the only caveat being that if you wanted the Pear network and the relay network to be one global community, you specifically need bridge nodes.

Also just a note; After bootstrapping your peer network a little, you don’t actually have to be online at the same time. Because your peers all have a “blind forwarding” cache of messages that they will pass around until the other user comes online again. So even my one-to-one DMs on Keet or any other Pear app still work fine for small message transfers because *any* of my peers online will hold and forward the message automatically. Kind of incredible how well it works. I haven’t even thought about missed messages for quite some time using the app. After you have even just a couple of peers using it with you, it’s shockingly reliable.

yes, relays are like POP in the sense that they cache messages... teasing out the differences from simple cache to authoritative archive is a separate matter, i'm working on this in my every day, can't stop thinking about it of course - we are turning an altcoin system into the archive... my boss doesn't quite fully get it either, he was talking up the idea of relays broadcasting and i was sitting there quietly thinking to myself (dude, bro, we are selling them the idea of their system being the archive)

this is much as i am pitching and working towards in my own work - nostr is the communication layer for everything

i am doing my best to insert it into an altcoin network as a pubsub for social networking, they seem to be getting it that it's neutral and allows them to deploy stuff with a neutral medium

it's really hard to say what it is explicitly, most normies need some cloak around it to get it, but they all get it when you know what their delusional threat model is

fair enough, and points taken.