What would prevent someone from namesquatting all the popular names?
Discussion
yeah, my name service consensus design doesn't include a mechanism for slowing down initial registrations, i need to think about how that should work.
the challenge is how to do it without mixing a binding to bitcoin with it and polluting the chain with anchors. i'd prefer not to do that. could just be a consensus rule that nodes will only accept the first one they see, and then they wait some period, disregarding all submissions until a new epoch for registration. something like rate limiting using a distributed consensus.
squatting is already handled in the protocol i designed in a similar way. names are just registered for a year only, and in the last 30 days of the registration the registrant has the right to extend it, so they are forced to be active in maintaining their squats, in a way that requires a lower time preference. there could be better methods to avoid this, but this is a simple one that doesn't involve monetary tokens.
Initially the 1, 2, 3 letter names could be reserved for internal use and for the 4 letters and up a donation of a certain amount of sats could be asked inversely proportional to the length of the name.
For example for 1,000 sats someone could get a 10 letters name.
And up with the same logic to get to 1 full BTC or more for 5 letters names, this could be also a smart idea to fund the project : )
Claude.ai suggested me earlier today also other possible attack scenarios:
Username Squatting:
Attacker registers thousands of popular names
Defense: Rate limiting, cost per registration, reputation system
Federation Collusion:
5+ federation nodes collude to register conflicting usernames
Defense: Bitcoin timestamp provides proof of fraud, community can fork
51% Attack on Bitcoin:
Attacker reorganizes Bitcoin blockchain
Defense: Wait for deep confirmations (6+), extremely expensive ($billions)
NOSTR Relay Censorship:
Relays refuse to propagate certain registrations
Defense: Users can query Bitcoin directly, run own relays
DNS Hijacking (if using domain):
Attacker compromises domain used for federation
Defense: Use IP addresses, Tor hidden services, or multiple domains
This is a hurdle that appears over and over again in decentralized systems: a solution that appears simple at first rapidly explodes in complexity. Which means the space of all possible protocols to implement the basic idea becomes very large. But the system only works if a large community can come to agreement on all of the details of the protocol. In the absence of a Schelling point, what’s the mechanism to get everyone to use your preferred version of the protocol over some other version?
The only solution I’ve ever come up with is WoT. First, make a decentralized list of protocols, using the simplest possible system, which is what I’ve designed Decentralized Lists to be. Next, curate the list using methods as proposed in the NIP. Lastly, hope that my WoT and your WoT and most people’s WoT will spit out the same (or almost the same) curated list. This last part I call “loose consensus” and I think it will work surprisingly well. But no one has demonstrated “loose consensus” — the act of demonstrating loose consensus would be an important milestone in the development of WoT on nostr. It might be an essential ingredient in preventing solutions like the one we’re discussing in this thread from falling apart.
I ought to write up an article on this; demonstration of loose consensus would make an excellent project for the #wotathon . (It would be an extra feature of a generic Decentralized Lists app — I do already have an article on that.)
yeah, the consensus doesn't tend to be the hard part, especially when you use WoT. but rate limiting is challenging because of the low cost of generating new keys. i'm doing a bit of a survey of mechanisms for rate limits in distributed systems that don't involve security deposits or proof of work. there must be something.
the complexity is something i also want to avoid, especially because the protocol design i have (it's based on a distributed database consensus called PnyxDB), is so simple.
oh, first thing springs to mind - to register a name, you must use it, a minimal number of times in the provisional period after registration. if the network never sees activity that reflects utilization, it can expire the provisional registration and it becomes open again, but the IP address and pubkey are both unable to attempt registration.
something like this.
i'm gonna start working on implementing this, this weekend, two things on my agenda for my downtime.
i also call it a "loose" consensus but actually the proper word is "weak consistency". bitcoin has weak consistency also, it is probabalistic. a WoT based consensus, like PnyxDB, where peers make trust attestations (or distrust) and distribute them so that you can use that graph to decide if that peer is near enough to your circle that it can be trusted.
oh yeah, if you never saw it:
https://github.com/technicolor-research/pnyxdb/blob/master/README.md
i've been following the guy who created this, handle is lesterpig, he is a distributed systems researcher, pnyx is the current revision of his protocol, previously he called it SporeDB.
distributed consensus designs that use Web of Trust are not common. i'd love to hear about others than this one that you know of.
Are people using pnyx?
I think my Decentralized Lists custom NIP would qualify as a distributed consensus design using WoT, if I understand correctly what you’re referring to. There are so many ways that this NIP could have been made more complex, but complexity is fatal if you want a large community to form around it. This is the simplest NIP I’ve been able to come up with and in my mind it is a primitive that can be used to build lots of more complex decentralized data structures (more complex that just simple lists).
i think the repo refers to this or even contains it but this is the paper on arxiv: https://arxiv.org/pdf/1911.03291
it is considered experimental, and both of our proposals are also. but i'm pretty sure there is something. i'm searching now for some exampes of production systems that are based on it, or use it in any way...
Production Systems
1. Stellar Consensus Protocol (SCP) - Federated Byzantine Agreement
Most significant web-of-trust-based consensus in production
Each node chooses its own quorum set (trusted validators)
Powers the Stellar network with billions in transaction volume
First provably safe consensus with: decentralized control, low latency, flexible trust, and asymptotic security
Anyone can join; no central authority dictates membership
Introduced by David Mazières as a novel trust model
2. Ripple's Unique Node List (UNL)
Each XRP Ledger server maintains a UNL - list of validators it trusts not to collude
Default configuration uses lists published by XRP Ledger Foundation and Ripple
Requires ~90% overlap between UNLs to prevent forks
Powers the XRP Ledger with significant daily volume
If nodes don't agree with their UNL validators, the network halts (safety-first design)
3. Proof of Authority (PoA)
Validators stake reputation and identity rather than computational power or tokens
VeChain is the primary deployment (101 Authority Masternodes)
Used by Walmart China, BMW for supply chain transparency
New blocks every ~10 seconds
Transitioning to DPoS in December 2025
Also used by: Bitgert, Palm Network, Ethereum testnets (historically Rinkeby/Kovan)
Popular in enterprise/consortium blockchains via Hyperledger Besu (IBFT 2.0, QBFT)
Research Prototypes
Multiple trust-enhanced Byzantine Fault Tolerance variants exist in research:
ModET-FC & TC-PBFT: For decentralized social networks
GABFT: Uses HonestPeer++ reputation model
TV-BRAFT: Trust values integrated into Raft
PoTC (Proof-of-Trust Collaboration): For M2M/IoT applications
Key Insight
While traditional consensus algorithms (Paxos, Raft, classic PBFT) dominate general distributed systems, web-of-trust mechanisms have found significant adoption in blockchain/DLT contexts where decentralized trust is essential. SCP and Ripple's UNL are the most mature implementations with real economic value at stake.
-----
i know one of the people who were working on VeChain back in the day, but i'm very hostile to monetary token based solutions. however, the info is clear, there is systems already running in production for a long time that have a strong WoT based component in them.
so even though it's "experimental" if the protocol seems like it might work, we should try to run it and see if we can at least get a network with a few dozen nodes or more and see how it works. nobody can use a system without a system to use.
the fact stellar uses a similar thing with assigning trust values to validators seems very pertinent. this network has managed to keep running for a very long time. it hasn't succeeded in acquiring a large userbase, however. but it clearly is secure enough that the protocol is still operational.
oh yeah, Steem (and hive) are specifically social network related but they are proof of stake, trash from that Dan Larimer Bytemaster guy. but their consensus is stake based, the stake is delegated by users voting for them.
i think it's worth taking a shot at building one and since you and I both want to do such a thing, it would probably be productive to continue a dialog, if you are interested, as it really depends on enough users to actually prove the concept.
I def want to build a system of distributed consensus using WoT — and a neo4j nostr relay will be a very useful step in that direction!
If you like, I can propose a schema for this. (Brainstorm doesn’t need a fully functional neo4j relay — it can make do with a much more restricted schema as described in my docs, at least for now — but a fully functional relay would be awesome and is def worth doing. Nothing like going above and beyond!)
Cool. Will work on it probably tonight and let you know when it’s done.
I can see two ways to go about it.
The first way would be to store the entirety of the data, including content fields, inside neo4j. I guess this would be the “purist” way, but events with lots in the content would eat up memory.
So the second way would be to omit content (and sig) properties from neo4j (with perhaps some exceptions, like kind 0 events), the goal being to reduce the amount of memory required for neo4j. In this method, we would view neo4j as an adjunct to a non-graph database, which could be relational, KV, etc, which would store each event in its entirety.
Perhaps I’ll design a schema for the purist method first and we can switch to the second method if/when bloat becomes too much of a problem.
https://git.nostrdev.com/mleku/next.orly.dev/src/branch/main/docs/FIND_RATE_LIMITING_MECHANISMS.md
this lists a number of options that have been used somewhere or discussed previously. a few of them might work. as i see it the main thing i haven't covered in the design is that initial bonanza problem.
yes, also, name->npub registration has been added to my design, in addition to all the standard DNS records. the idea would be that aside from the connectivity to the data source, it would be functionally equivalent to BIND
I share your hostility towards monetary token based solutions. We need a WoT system that people will use and that does not rely upon shitcoinery.
btw I sent you some DMs (from Damus) — not sure if they’re going through; if not should I use coracle maybe?
coracle properly supports nip-44, and it already worked before with the nip04 messages we have exchanged. so, yeah.
Or another way to self finance could be that a certain amount of "shorter names coupons" could be given out as reward for example for people and orgs running the federated relay system.