Going over zero sync from the replicache people

I accidentally had a coherent moment on stream today, so I made a video about it.

[AaronBoodman]

Hey thanks for the review. To answer your questions:

- zero is kind of like firebase, but fast, better api, and most importantly: works with your existing backend database. IOW zero replaces REST (or GraphQL)
- your backend is dramatically reduced. It can still exist and can be used for custom (non-crud) mutators, but it’s not needed on the read path.
- zero isn’t React-specific, it can work with any js-based frontend.

[marccawood]

I wonder how the client can know, if I issue a query, that it has ALL the data I want. It may have 10'000 rows but it can't know that 10'001 has appeared in the central server DB without querying the server. And if it does this, and calculates the relevance of the row for the current query on the server - then the cache is irrelevant. If it does it on the client that will be bad performance - I assume, much worse than a high powered DB. I guess the answer is: just try it, it works, it's blazingly fast.

[abdullahmomoh]

This really got me asking a bunch of questions. How does the central store for synching and persisting the data, in case it all goes down gets stored? Also how are things like webhooks going to be consumed, serverless?