#P2P projects keep failing socially because adoption is tiny. The #Fediverse succeeds socially because it keeps social #UX familiar. The path forward is a half-step strategy: bridge #fediverse + #p2p in real, usable ways until decentralised clients are socially relevant.

We need: Bridges & killer apps, seamless UX that makes federated + p2p content feel like one stream. A server that reads from both channels without making the user care about protocols.

A. what is happening with protocols:

* The #nostr crew are the children of #web3 mess, they are a bit reformed, let’s see.
* Then the #BlueSky are the reformed children of the #dotcons
* The #fediverse is the child of the #openweb
* #dat is a child of the #geekproblem if it is reformed or not, you can maybe tell me?
* #SSB was a wild child, now sickly/lonely with the #fashionable kids gathering round #nostr
* #p2p was the poster child of the era of the #openweb it was caught in the quicksand of legal issues, the shadow that was left was eclipsed by “free to use” #dotcons Now finds it hard to come back due to mobile devices not having an IP address, thus most people not actually able to use p2p reliably.

Q. ssb has technical shortcomings. It cant sparsely replicate data and verify it. It needs to download all data ever created by a user to verify, which makes it infeasible for many use cases. The main underlying data format is also hard to fix and leads to performance bottlenecks. The main founder moved on and it seems most ssb people are also looking for a new home.
dat’s time has not yet started as it approached things from a much more fundamental perspective. The initial vision was “git for any kind of data”, which means “version control for any kind of data” (peer to peer). The stack only now reached maturity to build proper tools on top of it. You have the dat-ecosystem with 2-3 dozen projects.
You have the holepunch/pears project which built a phenomenal “never on a server” desktop/mobile p2p video conferencing messenger with built in file sharing.
The app works flawless on mobile and is called https://keet.io
Also https://dat-ecosystem.org just now released it’s new website.
The https://pears.com runtime will be live in 5 days from now on the 14th of February for anyone to start hacking on p2p apps and some time later, the plan is to integrate it into the dat-ecosystem website, so anyone can start using p2p from within dat-cosystem page (which is an open source static website anyone can fork to get to the same) …no back ends required.
pears 🍐will only start working on the 14th of february. You can set a reminder.
The revolution starts then 🙂

A. will have a look, there are a few new #p2p projects reaching use at mo – the issue is none of them link to each other and likely thus non inter-op. This is the #geekproblem

Q. I don’t think there are any mature projects out there other than dat and ipfs. The former made by open source devs, self funded with a bit of help from public funding bodies, while the latter is the poster child of venture capitalists and got gazillions from investors. It’s the “big tech” of p2p.
Then you have a few less general purpose p2p projects which popped into existence in the last few years, but both dat and ipfs go back all the way to 2013 and it takes a lot to get things smooth and stable and support all use cases and get enough critical adoption and nodes to make the p2p network work.
That is why dat-ecosystem has a lot of existing projects that work and why it is reliable to build on top of it.
I do think the new more recent p2p projects in research state might become mature as well, but it will easily take them a few more years.
Many of those newer projects have people working on them part time only or focus on really special use cases and only time will tell if their approaches will bring something new to the table or not.
2024 will definitely be the year of dat, especially after February 14th, when pears.com goes live. This has been years in the making.
What started 2013 as (git for data) will now finally become it’s own independent p2p runtime. Goodbye nodejs & co. …and soon goodbye github & npm 🙂

A. https://holepunch.to/ its a very sparse website with no company info or #4opens process – it looks and feels like meany #dotcons if these projects do not link to each other or inter-op then they will fail like the hundreds I have seen fail over the last 20 years of this mess making. it’s a problem we can’t keep doing this shit, but we do. #4opens is a shovel to help compost this, can you do a write-up for these projects please.

Q. dat-ecosystem is a 501c3
It’s Code for Science and Society
And it is https://opencollective.com/dat
And it is governed by a Manifesto.
It is all on the website next to the “Info” button in the upper left corner.
If you mean pears.com ….that will change on February 14th
I didn’t mention holepunch.
Holepunch is just one of the many dat-ecosystem projects.
It is special, because one of the core developers of dat started it after he got a lot of funding and is currently maintaining many of the important code that powers dat and the dat-ecosystem projects.
But it doesn’t matter too much. The stack is open source under MIT and Apache 2.0 License for anyone to use. If holepunch would ever decide to stop maintaining the stack (which we do not think), dat-ecosystem can find other maintainers.

A. they are the owners of https://keet.io always look for ownership in #dotcons 🙂 a few of the ones I have been looking at over the last few years https://www.eff.org/deep…/2023/12/meet-spritely-and-veilid and the was a another one funded by NLNET they recently whent live, but can’t find the link. None of them link or interop, not even bridges. This is the #geekproblem

Q. Spritely is a great project.
It embraces the ocap security model (Object Capabilities).
It does apply it in lisp/scheme, which is a great fit with GNU Guix.
Their foundation is led by Randy Farmer.
Randy Farmer co-created Habitat with Chip Morningstar (an MMORPG) in the 1980s.
Chip Morningstar works with Mark Miller (Mentor of Christine Lemmer Webber).
Their project is called “Agoric”, which is a blockchain projcet funded by Salesforce.
They have their own Token and build a “Market Place”.
They as well work with ocap security model (but in JavaScript).
The JavaScript ocap version is what is known as SES and Endojs.
They regularly talk to make sure things are interoperable.
Ocap security is also what dat-ecosystem is embracing to pair it with peer to peer and bring it to the post-web. A version of the web not dominated anymore by big tech and big standard bodies.

#Veilid is a young and interesting project as well with a focus on anonymity over performance. This is a great use case that needs support, but dat was always about performance and any size of data and anonymity and privacy at all costs.
I’m not saying that is an unimportant use case, but there are plenty of solutions for extreme cases where anonymity and privacy are at utmost importance.
What is vastly more important imho is to have a p2p technology able to replace mainstream big tech services such as youtube, facebook, instagram, tiktok, google & co. because it won’t help us if we have a special niche technology that cant actually tackle big tech and surveillance capitalism but gives people some way to hide from it. …we need it too, but we also need a foundation on which to actually outcompete big tech imho.

Keet is a closed source peer to peer messenger & video conferencing app (might be open source in the future) and it is built on top of the dat stack.
The dat stack is very modular and in it’s core consists of a few main modules.
– hypercore, hyprebee & hyperdrive
– hyperdht & hyperswarm
– autobase
Those modules are maintained by holepunch, an organisation started by one of the core dat developers afte rreceiving a lot of funding to develop keet and now the pear runtime, which will be open source and public under https://pears.com after February 14th 2024 (Valentine’s Day ❤)
Keet itself is one of many apps, all part of the dat-ecosystem.
Most projects are open source, but not all, but they are all built on top of the MIT/Apache licenses p2p stack, which started as `dat` in 2013 and matured many years ago. The stack is battle tested and really works.
Of course – we all want everything open source and one day we might find a model, but if some closed source apps help bring in funding, it benefits the open source core.
Basically, you can think of “keet” as some fancy UI/UX on top of the open source software stack. Now sure – would be sweet if the UI/UX was open source as well, but then again, it’s not essential and until we transition into fully automated luxury Communism or whatever else works, something pays the bills and enables the open source core to be maintained 🙂
At least it works without any “Cloud Landlords”.
No servers, never on a server. No more cloud lords, a.k.a. Big Tech or #dotcons

A. The best we have currently is #ActivityPub DIY federated – this is community based (but fails in code to actually be this) which in meany ways is complemtery to #p2p based approaches – they are better together and if the can bridge or interop this is MUCH better, the #OMN is native to this.

Q. Yes. dat is very low level.
It would be cool to see somebody implement an activity pub based tool on top of it.
One dat-ecosystem project did it for nostr, but no activity pub yet.
I’m personally more interested into a desktop, terminal, version controlled data and software packages. “Social” tools are just one type of tools to built on top of the more fundamental p2p network and p2p system infrastructure.
I do think dat is good for laying these foundations, but “social” tools are a layer that dat as a stack will probably never focus on, but instead dat-ecosystem projects will hopefully take on that challenge 🙂

A. Some people are community based federated (the start of this conversation) others are individual, the #p2p world you talk about. This is not a fight they are both valid. As you say what we don’t won’t is more #dotcons 🙂 Good conversation on the state of #p2p I used to be much more involved in this side, but it failed with the move to #dotcons so got re-engaged when ActivityPub came alone the rebooting of web 1.5 😉 are you happy for me to copy this to my blog, can credit you or just use AQ anonymous format?

Q. any way you want. I dont think p2p has failed.
the p2p of the past was naive kids playing and it took a decade of adults and all the law enforcement they had at their disposal to bring it down and despite that torrents still run and even the piratebay continues to operate, although heavily censored.
Back then it was a few devs and a majority of users.
This time p2p is back and will enter mainstream open source developers after February 14th 2024 (5 days now).
This empowers an entire generation and anyone who wants to dive into p2p to build any kind of tool.
What was once hard and reserved to a few will be available to everyone.
We might see another nodejs/npm movement.
It loads a bit slow, but load this and check “all time”
This is the largest open source ecosystem humanity has ever experienced. http://www.modulecounts.com/
And while npm/github have been hijacked by microsoft, we will claw it all back soon
Btw. regarding Spritely and the backstory behind OCap, even though extremely technical in description, here is a summary of the work by Mark Miller et. al.
https://erights.org/history/index.html
https://en.wikipedia.org/wiki/Mark_S._Miller
> Miller has returned to this issue repeatedly since the Agoric Open Systems Papers from 1988
Mark Miller is Christine Lemmer Webbers Mentor.
He works with Chip Morningstar (who with Randy Farmer did Habitat in the 80s)
Randy Farmer is Executive Director of the Spritely Institute.
Agoric is the Cosmos Framework based Blockchain now.
https://agoric.com/team

A. Interesting to look back at all this stuff, reminds me I had dinner with https://en.wikipedia.org/wiki/Ted_Nelson in Oxford 20 years ago, he was a little eccentric with a clip on digital recording device, every convention had to be record. good to catch up with history https://www.youtube.com/watch?v=h-t405_JAJA that is more relevant today.

Q. Yes – peer to peer is hard. Not as a user, it is actually easy enough, but as a developer. Building p2p is not taught anywhere and there aren’t online learning resources the same way you can learn how to set up your react app, etc…
This will change after February 14th 2024 when the pears.com runtime is released. It is powered by the same p2p stack that developed with dat since 2013.
If anyone of you is a developer or has friends who are, you are all invited to dip your toes into the dat water 😛 …and start a new p2p project and join the dat-ecosystem 🙂 It will get quite easy in 4 days from now and it will again get a lot easier in the coming weeks when more examples and docs are publishes and others build as well.
The Storyline around Mark Miller, Randy Farmer & Chip Morningstar is totally separate from it, but it is also important, because it is what powers
1. the Spritely project and Christine Lemmer Webber
2. the Agoric Blockchain Project backed by Salesforce
3. the Ethereum Metamask Wallet and Co.
It also influences the big standards bodies and I see it two fold.
It’s a story about philosophy, values and vision driven by the specific people in it.
It is also a story about “object capabilities” which is a powerful perspective on security and will enable and inform a lot of p2p interaction which without would require some sort of centralized servers, but with ocap can do it on it’s own p2p

A lightly edited conversation between Hamish Campbell (A) and Alexander Praetorius (Q)