CRDT hacker here. The talk says CRDTs need to store a history of fine-grained edits indefinitely to make them work. There are some ways around this requirement. I'd be happy to chat more about this if anyone is interested!
I’m new to CRDTs, but could we clean up tombstones if all replicas acknowledge? Not too sure, but could we persist snapshots of state at certain “compaction” points?
You can do this by (1) extending your CRDT into a CTM (see https://braid.org/time-machines) and then (2) use the antimatter algorithm (braid.org/antimatter), or something similar, for acknowledgements.
The antimatter algorithm allows peers to learn where the rest of the network has caught up to, without a central actor, or relying on consensus, across arbitrary P2P connection & disconnection events in the network. It even allows subnetworks, after a partition, to prune the history that they generate while partitioned, while still holding onto the necessary older history to reconnect with the other half of the partition.
Yep. As well as what Toomim said, eg-walker (another text crdt algorithm) doesn’t need or use tombstones. With eg-walker we just store the history of all changes (kinda like git). Some changes are deletes. The changes are only used for merging old stuff, so if you’re happy to set a cutoff (or have some coordination like antimatter) you can delete the operation history and all the overhead that goes along with it.
"Centralised systems were designed with the best of intentions, but were turned against us anyway."
What a weird take. The internet was built fundamentally decentralized but was centralized against us with the worst of intentions. They lost me at the first sentence.
I think it's a bit of a stretch to include protocols and protocol suites among centralized services. One simple test for this is the question: "How many Xs are there?". For examples, how many email servers, FTP servers or terminal servers are there? Compare that with "How many Facebooks or GitHubs are there?".
Email protocol suite is designed to be federated. FTP is just a file system access protocol. But you could combine it with an inter-server filesystem synchronization protocol/service to make it a distributed federated service. And as for terminal servers,.. well, I don't think centralization makes much sense there. How can you achieve any of these with centralized services?
I talk about the past, your university FTP-Server was the central point to get your Software/Manuals also publish your work (today's Github/Sourceforge?). Your university Email Server was the primary central point to exchange Information mostly inside your university.
Again i talk about the past when email was primarily used to talk to other peoples often not even over a net but inside a mainframe thing.
I though i was clear talking about the past hence not including Facebook or GitHub, and btw. Email just became "federated" when everyone agreed to use smtp when talking over the internet.
It really boils down to if the individual cares about freedoms. Some hackers don't care about them, and some normies do. Both of them use centralized and restrictive services. But the former does it by choice, while the latter does it because they don't know any better. But those normies do take an action when they have enough information. How many ordinary people have participated in boycotts and cancelation of subscriptions against corporations in protest of exploitation or for digital detoxification?
It's partially our own failure to be loud enough and get them the information they need.
Firefox was always a geek thing, I've been using it since it was called phoenix and at its peak, it was mostly nerds installing that at schools, convincing their mums, etc, thanks to the adblock.
The only reason we are not all using edge is because google spend billions marketing Chrome in early 2000. They got the normies with brute forcing, because they could make money with it, not for making the world better.
Heck, they promoted google on the HOME PAGE of google search, an ad spot you can't even buy from google, with a pseudo notification, a format google uses for nothing else in ads.
They went full throttle.
But nobody is going to spend millions to promote decentralization. Because it's about concentrating less power.
HN has always been terrible at undertanding that, because while they argue about what browser to use, the average user can barely make the difference between an app and a website anyway.
>I've been using it since it was called phoenix and at its peak, it was mostly nerds installing that at schools, convincing their mums, etc, thanks to the adblock.
Are you talking about Netscape? Because that was installed on everything ;)
>because google spend billions marketing Chrome in early 2000
Open systems provide plenty of opportunity for smaller businesses to provide innovative and convenient services. What kind of innovation and experimentation other then getting more ads in are the gatekeepers that we have today interested in doing as long as there's no competition?
CRDT hacker here. The talk says CRDTs need to store a history of fine-grained edits indefinitely to make them work. There are some ways around this requirement. I'd be happy to chat more about this if anyone is interested!
I’m new to CRDTs, but could we clean up tombstones if all replicas acknowledge? Not too sure, but could we persist snapshots of state at certain “compaction” points?
You can do this by (1) extending your CRDT into a CTM (see https://braid.org/time-machines) and then (2) use the antimatter algorithm (braid.org/antimatter), or something similar, for acknowledgements.
The antimatter algorithm allows peers to learn where the rest of the network has caught up to, without a central actor, or relying on consensus, across arbitrary P2P connection & disconnection events in the network. It even allows subnetworks, after a partition, to prune the history that they generate while partitioned, while still holding onto the necessary older history to reconnect with the other half of the partition.
Yep. As well as what Toomim said, eg-walker (another text crdt algorithm) doesn’t need or use tombstones. With eg-walker we just store the history of all changes (kinda like git). Some changes are deletes. The changes are only used for merging old stuff, so if you’re happy to set a cutoff (or have some coordination like antimatter) you can delete the operation history and all the overhead that goes along with it.
Very inspiring talk how to organize data for federalized systems.
worm-blossom crew is just a delightful bunch of humans doing really great work
http telnet, irc, SMTP, imap - what else we need really? KISS
Encryption is awful for future archeologists etc.
"Centralised systems were designed with the best of intentions, but were turned against us anyway."
What a weird take. The internet was built fundamentally decentralized but was centralized against us with the worst of intentions. They lost me at the first sentence.
You have to go into the past a little bit. Think about your:
University email, FTP, and terminal server.
The Internet is just a highway. You will end up at a destination.
I think it's a bit of a stretch to include protocols and protocol suites among centralized services. One simple test for this is the question: "How many Xs are there?". For examples, how many email servers, FTP servers or terminal servers are there? Compare that with "How many Facebooks or GitHubs are there?".
Email protocol suite is designed to be federated. FTP is just a file system access protocol. But you could combine it with an inter-server filesystem synchronization protocol/service to make it a distributed federated service. And as for terminal servers,.. well, I don't think centralization makes much sense there. How can you achieve any of these with centralized services?
I talk about the past, your university FTP-Server was the central point to get your Software/Manuals also publish your work (today's Github/Sourceforge?). Your university Email Server was the primary central point to exchange Information mostly inside your university.
Again i talk about the past when email was primarily used to talk to other peoples often not even over a net but inside a mainframe thing.
I though i was clear talking about the past hence not including Facebook or GitHub, and btw. Email just became "federated" when everyone agreed to use smtp when talking over the internet.
Not against us. Against some of the nerds will. The rest cheered. That's why it worked.
Humans hate friction, they don't want to pay for maintenance and have short term thinking.
Even on HN there are plenty of voices saying they won't even bother using firefox because it inconvenience them.
Can we blame then the normies for choosing integrated easy systems to use?
It really boils down to if the individual cares about freedoms. Some hackers don't care about them, and some normies do. Both of them use centralized and restrictive services. But the former does it by choice, while the latter does it because they don't know any better. But those normies do take an action when they have enough information. How many ordinary people have participated in boycotts and cancelation of subscriptions against corporations in protest of exploitation or for digital detoxification?
It's partially our own failure to be loud enough and get them the information they need.
>Can we blame then the normies for choosing integrated easy systems to use?
With that logic everyone would use the Edge-Browser right? Don't underestimate the "normie" ;)
Firefox was always a geek thing, I've been using it since it was called phoenix and at its peak, it was mostly nerds installing that at schools, convincing their mums, etc, thanks to the adblock.
The only reason we are not all using edge is because google spend billions marketing Chrome in early 2000. They got the normies with brute forcing, because they could make money with it, not for making the world better.
Heck, they promoted google on the HOME PAGE of google search, an ad spot you can't even buy from google, with a pseudo notification, a format google uses for nothing else in ads.
They went full throttle.
But nobody is going to spend millions to promote decentralization. Because it's about concentrating less power.
HN has always been terrible at undertanding that, because while they argue about what browser to use, the average user can barely make the difference between an app and a website anyway.
>I've been using it since it was called phoenix and at its peak, it was mostly nerds installing that at schools, convincing their mums, etc, thanks to the adblock.
Are you talking about Netscape? Because that was installed on everything ;)
>because google spend billions marketing Chrome in early 2000
Really? Early 2000?
No, firefox was first called phoenix (because it was seen as a rebirth of the netscaoe ethos) then firebird.
Netscape got killed by ie 5.5 and firefox competed mostly with ie 6 on Market and opera on innovation.
And yep, i think 2008 was first chrome ad on google search. It hurts but we are a quarter century in now.
>And yep, i think 2008
That's late 2000
Firefox was installed on around 30% in 2010:
https://statisticsanddata.org/data/most-popular-browser-1996...
Open systems provide plenty of opportunity for smaller businesses to provide innovative and convenient services. What kind of innovation and experimentation other then getting more ads in are the gatekeepers that we have today interested in doing as long as there's no competition?
I'd read that as "some centralised systems", not "all centralised systems"
[dead]