alt Hacker NewsThere are invariably 3 solutions to all “federated” web problems:
- Semantic HTML
- Web indexing services
In this case, it's mostly indexing and a little of the other two if you want to add social features. I despite ATProto with all my soul for taking a problem with such a simple and standard solution and totally obscuring it behind hundreds of layers of JSON, faux federated services, and technical jargon, all in service of creating an inferior version of Twitter. I guess it wouldn't be as sexy to offer a web indexing service instead.
Genuinely curious, how would you build it differently? Would you stick with atproto's indexing model and try to simplify it, or would you use another approach?