Oddbean

Architecture is a tricky thing. Very few people know the history of the URL aka the URI. The "U" stands for Uniform. But Tim wanted to stand for Universal. Why is this a big deal? Well, Tim wanted the web to be a universal decentralized space, with http just being one of the possible prefixes of protocols. He wanted it to include ftp, file system, email, any protocol you can imagine (e.g. including nostr). All through a huge "web" of hyperlinks. The IETF had another idea. They wanted a massively scalable, low-latency network that could scale to the planet and deliver huge amounts of information to the edge, very cheaply. The IETF won that battle, and Tim says he regets not pushing harder. In face, the IETF were right. Once you scale something it opens to doors to every other possibility, not the other way round. What's that got to do with nostr and architecture? Well we have exactly the same problem with nostr architecture. From a very promising start in the first 2 years, little progress has been made on creating a scalable uniform interface and infrastructure that nostr can grow into. Even if nostr had an influx of users, it's not clear whether the neglected relay network can handle an OOM more. You often hear stories of Damus relay struggling, and rarely hear stories of web systems or bluesky struggling. Nostr through it's taproot identity has the "universal" nature that Tim likes, but it lacks a scaling architecture or infrastructure that has seen nostr's rivals start from small and grow 100x bigger. Yes, we do need great clients, search and content, of course we do. But this dual nature of architecture ie features vs scale is an area where more work is needed. Creating a uniform architecture is going to be really hard, because it's hard for people to agree. Alex Gleason seems to be the person that is looking at this, in the most serious way, so far, imho.