datahike

whilo 2026-03-25T08:54:02.266119Z

New blog post on our memory model, feedback welcome! https://datahike.io/notes/collaborate-without-infrastructure/

2026-03-25T10:08:14.357549Z

Interesting read, thanks for sharing.

whilo 2026-03-25T10:14:42.256539Z

@ray1729 You are welcome! Anything that was unclear or could be improved?

2026-03-25T10:17:56.019139Z

I thought it was very clear. The only thing that caught me out was the jump from talking about browsers to the cross-database join example, I was expecting a browser example and had to do a double-take.

whilo 2026-03-25T10:21:19.911449Z

Thanks, that makes sense. Yes, that could probably flow better.

rschmukler 2026-03-25T15:52:57.430689Z

I was doing a little research on branching in datahike. If branches are implemented as different root pointers does that mean that you could have two writers concurrently, so long as each one is just writing to each of their own branches?

rschmukler 2026-03-25T17:19:10.973769Z

With the in-memory store this seems to work fine across multiple writers. Would it hold for other stores?

whilo 2026-03-25T18:49:26.189089Z

Yes, you effectively have a connection and writer per branch already. It should work with any store.