nostr has no global source of truth, and that is a good thing
Out of interest, I follow the progress of a lot of other projects similar to nostr, and a couple links surfaced today:
BlueSky has a big "firehose" connection that streams all updates (new posts, reactions, etc) to subscribers. Unsurprisingly, this is difficult to process except on beefy servers with lots of bandwidth. So, one proposed solution is to strip out all that pesky cryptography (signatures, merkle tree data, etc):
Jaz’s Blog
Jetstream: Shrinking the AT Proto Firehose by >99%
Jetstream is a streaming service that consumes an AT Proto Sync Firehose and converts it into lightweight, filterable, friendly JSON, allowing us t...
And over on Farcaster, keeping their hubs in sync is too difficult, so they want to make all posts globally sequenced, like a blockchain. The details are still being worked out, but I think it's safe to assume there will be a privileged global sequencer who decides on this ordering (and possibly which posts are included at all):
GitHub
FIP: Introducing Ordering · farcasterxyz/protocol · Discussion #193
Problem Farcaster hubs are finding it harder to stay in sync. The problem is that user data changes at a very rapid pace. If you aren’t familiar ...
In my opinion, both of these issues are symptoms of an underlying errant philosophy. These projects both want there to be a global source of truth: A single place you can go to guarantee you're seeing all the posts on a thread, from a particular user, etc. On BlueSky that is
https://bluesky.app and on Farcaster that is
Farcaster
Farcaster
A decentralized social network
.
Advocates of each of these projects of course would dispute this, pointing out that you could always self-host, or somehow avoid depending on their semi-official infrastructure, but the truth is that if you're not on bluesky.app or warpcast.com, you don't exist, and nobody cares that you don't exist.
nostr has eschewed the concept of global source of truth. You can't necessarily be sure you are seeing everything. Conversations may sometimes get fragmented, posts may disappear, and there may be the occasional bout of confusion and chaos. There is no official or semi-official nostr website, app, or relay, and this is a good thing. It means we are actually building a decentralised protocol, not just acting out decentralisation theatre, or pretending we'll get there eventually and that the ends justify the means.
Back when computers were primitive and professional data-centres didn't exist, it was impossible to build mega-apps like Twitter. Protocols had to be decentralised by default -- there was simply no other way. We can learn a lot by looking back to protocols of yesteryear, like Usenet and IRC, and still-popular protocols like email and HTTP. None of these assume global sources of truth, and they are stronger and better for it, as is nostr.
Replies (56)
and it has a cool name ring - noString :-)
Really.
"It means we are actually building a decentralised protocol, not just acting out decentralisation theatre"
Doug Hoyte
nostr has no global source of truth, and that is a good thing
Out of interest, I follow the progress of a lot of other projects similar to nostr, and a couple links surfaced today:
BlueSky has a big "firehose" connection that streams all updates (new posts, reactions, etc) to subscribers. Unsurprisingly, this is difficult to process except on beefy servers with lots of bandwidth. So, one proposed solution is to strip out all that pesky cryptography (signatures, merkle tree data, etc):
Jaz’s Blog
Jetstream: Shrinking the AT Proto Firehose by >99%
Jetstream is a streaming service that consumes an AT Proto Sync Firehose and converts it into lightweight, filterable, friendly JSON, allowing us t...
And over on Farcaster, keeping their hubs in sync is too difficult, so they want to make all posts globally sequenced, like a blockchain. The details are still being worked out, but I think it's safe to assume there will be a privileged global sequencer who decides on this ordering (and possibly which posts are included at all):
GitHub
FIP: Introducing Ordering · farcasterxyz/protocol · Discussion #193
Problem Farcaster hubs are finding it harder to stay in sync. The problem is that user data changes at a very rapid pace. If you aren’t familiar ...
In my opinion, both of these issues are symptoms of an underlying errant philosophy. These projects both want there to be a global source of truth: A single place you can go to guarantee you're seeing all the posts on a thread, from a particular user, etc. On BlueSky that is
https://bluesky.app and on Farcaster that is
Farcaster
Farcaster
A decentralized social network
.
Advocates of each of these projects of course would dispute this, pointing out that you could always self-host, or somehow avoid depending on their semi-official infrastructure, but the truth is that if you're not on bluesky.app or warpcast.com, you don't exist, and nobody cares that you don't exist.
nostr has eschewed the concept of global source of truth. You can't necessarily be sure you are seeing everything. Conversations may sometimes get fragmented, posts may disappear, and there may be the occasional bout of confusion and chaos. There is no official or semi-official nostr website, app, or relay, and this is a good thing. It means we are actually building a decentralised protocol, not just acting out decentralisation theatre, or pretending we'll get there eventually and that the ends justify the means.
Back when computers were primitive and professional data-centres didn't exist, it was impossible to build mega-apps like Twitter. Protocols had to be decentralised by default -- there was simply no other way. We can learn a lot by looking back to protocols of yesteryear, like Usenet and IRC, and still-popular protocols like email and HTTP. None of these assume global sources of truth, and they are stronger and better for it, as is nostr.
> Advocates of each of these projects of course would dispute this, pointing out that you could always self-host, or somehow avoid depending on their semi-official infrastructure, but the truth is that if you're not on bluesky.app or warpcast.com, you don't exist, and nobody cares that you don't exist.
Aren’t nostr adepts of mythical outbox who consider anything server side to be an evil centralization from “them” saying the same things? “No one guarantees relays keep your posts but you can always selfhost and no one will care about you and this is a good thing because Nostr is a pure signal and we don’t need stupid normies here!”
Decentralized ends require decentralized means.
🍻👋🧡
What a great post to not be able to Zap. 😤
Nope. Ordering is the killer feature of nostr. They just havent figured it out yet. But the competition has.
View quoted note →This note will also help
Doug Hoyte
nostr has no global source of truth, and that is a good thing
Out of interest, I follow the progress of a lot of other projects similar to nostr, and a couple links surfaced today:
BlueSky has a big "firehose" connection that streams all updates (new posts, reactions, etc) to subscribers. Unsurprisingly, this is difficult to process except on beefy servers with lots of bandwidth. So, one proposed solution is to strip out all that pesky cryptography (signatures, merkle tree data, etc):
Jaz’s Blog
Jetstream: Shrinking the AT Proto Firehose by >99%
Jetstream is a streaming service that consumes an AT Proto Sync Firehose and converts it into lightweight, filterable, friendly JSON, allowing us t...
And over on Farcaster, keeping their hubs in sync is too difficult, so they want to make all posts globally sequenced, like a blockchain. The details are still being worked out, but I think it's safe to assume there will be a privileged global sequencer who decides on this ordering (and possibly which posts are included at all):
GitHub
FIP: Introducing Ordering · farcasterxyz/protocol · Discussion #193
Problem Farcaster hubs are finding it harder to stay in sync. The problem is that user data changes at a very rapid pace. If you aren’t familiar ...
In my opinion, both of these issues are symptoms of an underlying errant philosophy. These projects both want there to be a global source of truth: A single place you can go to guarantee you're seeing all the posts on a thread, from a particular user, etc. On BlueSky that is
https://bluesky.app and on Farcaster that is
Farcaster
Farcaster
A decentralized social network
.
Advocates of each of these projects of course would dispute this, pointing out that you could always self-host, or somehow avoid depending on their semi-official infrastructure, but the truth is that if you're not on bluesky.app or warpcast.com, you don't exist, and nobody cares that you don't exist.
nostr has eschewed the concept of global source of truth. You can't necessarily be sure you are seeing everything. Conversations may sometimes get fragmented, posts may disappear, and there may be the occasional bout of confusion and chaos. There is no official or semi-official nostr website, app, or relay, and this is a good thing. It means we are actually building a decentralised protocol, not just acting out decentralisation theatre, or pretending we'll get there eventually and that the ends justify the means.
Back when computers were primitive and professional data-centres didn't exist, it was impossible to build mega-apps like Twitter. Protocols had to be decentralised by default -- there was simply no other way. We can learn a lot by looking back to protocols of yesteryear, like Usenet and IRC, and still-popular protocols like email and HTTP. None of these assume global sources of truth, and they are stronger and better for it, as is nostr.
Went to send you a zap, but says not setup… get on it sir!
Great note. Thanks. I like learning from people clued into the technical side
It’s not something that come naturally to me
As a commercial person… how did those older protocols make money?
Ie, is there a historical business model, or models, that we could mimic?
Best note of the day👑!
TCP/IP analogy was a real eye opener for me
It's the exact opposite. You can host your own relay, or use your neighbor's, and (outbox) Nostr clients will still care about you.
This website is censoring my phone :(
Try again, just zapped it.
this.
He doesn't have a lightning wallet connected. You will get your zap back in 24hours.
interesting... removing the merkle tree
tl;dr of the bsky / Jetstream article:
> If consumers already trust the provider to do validation on their end, they could get by with a much more lightweight data stream.
"the provider" is currently Bluesky. exclusively.
trust?
🫡
👏👏
Out of the 4 example you provided only Email could potentially be considered decentralized. Usenet may store articles in a distributed manner but you still have to connect to a central server to download a list of articles and the majority of articles in a Usenet network are stored on centralized servers, it is the contributions that are decentralized. That is why Usenet is so much faster that torrents, you pay the service so they can host and serve the files from beefy servers. IRC is entirely centralized, the only way to communicate on an IRC channel is to connect to the server it is hosted on. Same with HTTP, there is no way to access the web page without connecting to the central server. Even email is centralized, you could say it is federated but even federated services are just a distributed centralized system. Email is the only one of these services that you could argue doesn't have a centralized source of truth but I would argue that it still does.
I'm not saying decentralized systems are bad. Imo they are the only way the internet survives. However those examples aren't examples of decentralized systems.
A protocol isn't there, to make money.
The mistake is in thinking everyone _should_ care about everything and therefore everyone _must_ see everything. Outbox/Inbox ensures that you can see and interact with what is important to you, and ignore the rest.
Smaller relays gives people the ability to have access to a wide range of npubs, but focus on their own cluster and store their own data, by creating a system of hops.
I can move events from the Citrine relay on my phone to fiatjaf's personal relay, just by including an @ to his npub, in the note. It's like a mailing address. That's why we refer to the system as "mailboxes".
I don't need to know what else is on his relay. I would just need to see his response, sent to my address.
Yes just a bunch of leaves in bin for composting
So you expect everyone runs their own relay, just like everyone runs email servers today right?
Remember there is no centralization around gmail today — you can always selfhost since smtp is an open protocol. And if someone wont receive your emails well thats not a problem right?
"Back when computers were primitive and professional data-centres didn't exist, it was impossible to build mega-apps like Twitter. Protocols had to be decentralised by default -- there was simply no other way."
Ive been thinking about this statement today. Its why i am drawn to old unix, bsd, and plan9 systems and tools.
View quoted note →Everyone doesn't need to run a relay, but everyone could. Running a relay is already incredibly easy to do, and costs nothing, and that'll just get better.
It's not decentralized, if you need to traverse "approved" hubs to get from point A to point B. That's not only a protocol question, that's an implementation question.
We are emphasizing a complex, sprawling network of privately-run nodes, preferably communicating offline or over Tor, to avoid the centralized fate of email.
Every relay operator is free to let me use his relay or not. This is a nonviolent protocol. We do not need consensus in event content or originators. We just need lots and lots of different paths, and to ensure that anyone can build their own path.
Running relay is cheap and will only become cheaper and corporations are straight stupid paying millions of dollars monthly for infrastructure?
Or rather this “cheap outbox relay approach” ignores social media aspects and focuses on “private groups” scenarios which is more like my personal webpage no one knows about rather twitter alternative?
I also think you are underestimating the importance of the timing of introduction. When e-mail began to be popularized, in the mid-1980s, hardly anyone had a computer. So the few of us using it had it to ourselves and it was common to host an e-mail server because only computer geeks were on it (my dad is a network engineer).
Then it got really busy in the 2000s, but running your own server was a big technical hurdle, to most people, so they were happy to outsource the work to Google and etc. and that allowed for network consolidation, by creating email bottlenecks, where they would filter spam and etc.
Now, anyone can create their own, personal intelligent spam filter and run a relay on their smartphone or laptop, just by downloading an app. That's a game-changer.
It addresses both scenarios because you can write directly to the big ones.
You can explicitly choose the relay you want to write to, in some particular instance, while your outboxes tell others where a copy of your notes can _usually_ be found.
So, I am writing to a big relay, right now, but a copy is stored in my personal relay.
I still believe if everyone who is able to create their own personal intelligent spam filter does it (I stopped for example because don’t see any value in it right now) — all those people will be able to discuss are bitcoin, nostr, meat, anarcho capitalism, libertarianism and maybe some other cryptocurrencies, but unsure here. However, I define “twitter alternative” completely differently
Hope to be very wrong and your optimism to play out well 👍
The only reason why you can write to big ones right now is because no one uses nostr right now. There is no even theoretical incentive for public relays to operate if nostr really becomes something
And no one wants to address this issue. Instead people tend to over engineer 64K relays that are extremely efficient and can serve 256 out of a coffee machine
There is no direct incentive, for public relays, that's true. Incentives don't need to be direct. We run one for documents because we have developed a publishing/reader app. Facilitating the usage of the app is the incentive.
And we will have very large private relays, because of micropayments.
i can confirm about BS and FC, both are retarded... FC is actually less retarded than BC because they at least don't embed binary data nor use binary as their primary message format, but creating a global total ordering just turns it into another silo and it's pretty sad they don't get that
i'm working with Arweave on a project that involves both of these other protocols but recently i hear from them that they want to make arweave storage a fully persistent event storage for nostr and actually integrate properly with relays
different mindset, they are more interested in the fact that nostr is a delivery system instead of being stuck in the global consensus mindset... i also had problems with the Internet Computer Protocol people because their nodes have one replication setting which makes the cost of data storage absurdly high, and were "talking about" making lower reliability replication schemes for cheaper, larger storage... but who knows where that's gonna go
crypto project guys are completely mentally locked into the strongly consistent mindset and don't understand the most basic principle of distributed systems that you can only have 2 out of the three of CAP strong at the same time, and nostr has high availability and partition resistance (it's leaky, is how i describe that, it's not possible to stop replication across the network, partitioning is where you split the data) what nostr doesn't have is strong consistency, and that's the issue, they can have consistency, but the cost is either availability or partition resistance, and they are too dumb to realise that if they focus on consistency one of those two has to go
Let's put it this way:
Nostr is the only one of these "decentralized, open protocols" that is truly decentralized and open. Even the protocol spec itself, is completely open. I see no value in half-assing a communications revolution, so I can't even be bothered with the others. If this isn't successful, then damn it all. 😂
I guess that's a sort of optimism, yes. This is the only one crazy enough to actually work.
i think part of the reason why arweave people understand it better is they already have a separation in their architecture between replicas and the consensus and the delivery API service, so it is simpler for them to think of their database as storage, indeed that's one of the whole points of their project is pretty much focused on storage so they have more realistic strategy about how to keep the cost contained and scaling replication to fit the use case instead of having a one-consensus-to-rule-them-all mentality
also, the replication strategy of farcaster is patently retarded it's just IPFS
and they are not gonna have a good time trying to make a total ordering, this is precisely what you can't make work at scale!
Nostr is a subset of http
Relays are, yes.
That's like asking "how does internet make money?". It doesn't. Businesses built ON the internet can though 😉
Bingo. So what will work on Nostr?
Man I miss those times on IRC
Ça m'arrive souvent sur nostr et c'est fort dommage puisqu'il y a de ces notes qui mériteraient plus de zaps très sincèrement
Assez récurrent et fort dommage pour certaines notes
Doh! just realized, thanks.
Turn on zaps, my man!
Fort interessant votre note pour tout ce qui s'intéresse aux plateformes, leurs limites surtout de cette idée de s'assurer de qu'est est la *décentralisation * dans nos systèmes qu'on en dise demeurent centralisées jusqu'à présent même en gérant soit même des relais 😂🫂
C'est dommage mais une expérience d'apprentissage.
> is there a historical business model, or models, that we could mimic?
Good question! Usenet and IRC were often hosted by ISPs, which users paid for indirectly with their internet subscriptions. People also paid for access to Usenet. DejaNews was a popular archive/provider service before it was bought by Google. On IRC people pay for service providers to keep them constantly connected (bouncers) and provide vanity DNS names. Also, people pay to run bots like Eggdrop to manage/moderate their channels.
For email and HTTP there are the obvious hosting and other service providers, but I suppose the biggest value they support is the things built on top of them, which is maybe fitting for a general-purpose protocol.
Interesting perspective. We could debate what decentralised means, but I doubt we'd ever be able to find a universal definition -- it is too much of a spectrum. If your definition of decentralised is that there are no servers at all, then I guess you'd think only purely P2P protocols are, and not even nostr would qualify. From a total purity perspective, probably anything using DNS would be disqualified too.
My view on these protocols is as follows:
Usenet has essentially the same model as nostr. Yes, there are servers (relays), but people are free to choose which ones they use. They can post their messages to any of them, and those messages may get propagated to other servers. Each server can have its own message acceptance/forwarding policies, and choose which other servers to connect with.
IRC is also a decentralised network (the R stands for relay). An IRC network consists of many different servers relaying messages. Each server agrees roughly with the rules of the wider network, but is generally free to administer its server as it sees fit (including user bans, preventing relaying certain channels, etc). Sometimes server operators disagree, and this results in them leaving the network and establishing their own. That's why there are many different IRC networks, EFnet, IRCnet, DALnet, etc.
HTTP and email are both decentralised in the sense that you don't need to get anybody's permission to connect to the network, and there are no single points of failure.
we build a trustless world that we can all verify together ❤️
Usenet, IRC and nostr replicate data that's how they're decentralized. HTTP doesn't that's why it's so easy to censor.
Most value seems clearly created due to building atop Nostr. Which itself I think is very cool. A superior digital property right is a better foundation for capital creation
I like those services you mention eg ISP
