https://fossil.calmabiding.me/streamful/info/ea4d3ed020c368cd
This commit is quite thorough. It breaks down all the data structures and how one would produce a message that the streamful server could understand.
It also lays out some initial thinking about a simple aggregation use case. And I'm already starting to go down the rabbit hole of how I would want things to show up in the client for a very simple bluesky clone.
For example, how do you count replies?
If I wanted to model it properly, I'd have to keep track of the entire ancestry chain of a message. Which wouldn't be too terrible, I suppose! Maybe that will be the proof that aggregates aren't so hard to model.