The DAT project is sadly no more, but author Paul Frazee is now one of the lead developers at Bluesky.

It seems like it was so much better designed than IPFS, since it had a native concept of collections instead of trying to make every block independently accessible on the DHT.

I'm glad others recognise this problem too.

IPFS doesn't try to. Only top-level directories are added to the DHT. To fetch a child object, your node asks the same node it got the parent from.

Did they finally change that? Last I checked the default settings was to spew every block on the DHT unless you use roots mode, and to dump your entire list of wanted blocks to all 500 or so connected peers you often have, just in case maybe they have a block you want.

I thought Pear/Holepunch is the remaining spiritual successor from past Dat contributors?

https://docs.pears.com/

Last I checked much of their released stuff was still closed-source, not sure where they're at these days.

Could be, it's quite a while since I talked to Paul about it and we only discussed it briefly.

There's also Cabal.chat which works based on similar tech

I ended up reading through this whole line of holepunch stuff, and as much as I love almost all of it in idea, spirit, quality of work, the only way to run in browser or mobile is an entire REMAKE of nodejs they call "bare"

Sounds like a nightmare that would take forever. However all that gleam and bun and deno kind of makes it sound like its possible i guess.

I get sad when great tech is never adopted. What is the equivalent of civil war in the tech world where people need to be pushed so hard to consider changing anything? Its like the full government crackdown on eliminating encryption has to happen before we even start trying p2p

p2p is alive and well in the form of netcode

https://github.com/ThusSpokeNomad/GameNetworkingResources

You are spot on. We already have the technology. It's a bunch of idiots mucking around with stuff like IPFS.

If there was a real need for it, P2P adoption would occur very quickly.

Hi! Author here. Funny to see my term paper posted. A lot has happened in the space since then and I haven't kept up as well as I would have liked.

And more recent posts by different folks that you'd recommend reading instead?

Not that I know of :/ I think some stuff in the article applies broadly to content addressing and some is pretty implementation specific. I haven't kept up with the field.

(2018)

The beaker browser was so cool. Too bad that it never really rook off. Does anyone know if there are any similar projects to it?

Hey all, just letting you know that the core technology inside of Dat/Beaker is alive and well inside Keet [1] (a P2P communications app) and Pear [2] (a platform for building P2P desktop/mobile apps). Keet is built on Pear.

Both projects are built by Holepunch [3], a company we formed a few years back to keep the peer-to-peer candle burning. It's always a challenge but never dull. We're working on tons of mad science [4] all the time, and we're hiring [5]!

(Disclaimer: I work at Holepunch)

[1] https://keet.io/ [2] https://pears.com/ [3] https://holepunch.to/ [4] https://github.com/holepunchto/ [5] https://holepunch.recruitee.com/

+1 to this, if you’re interested in any of this work then give holepunch a look. It’s the same protocol engineers, same core tech, from the folks who worked on beaker.

When is it getting open sourced?

Lookup the hypercore protocol, https://github.com/holepunchto