One thing Firestore's JS client library does that is very interesting is that mutations by the in-browser client show up in persistent queries, even before they sync with the server, even offline. You can tell which changes are safe on the server and which not by metadata.
This greatly improves latency and simplifies code. Trying to manually merge the mutations into all of your queries would be very awkward.
Also, the whole concept of browser clients interacting directly with the database server is darn interesting. Combine that with better support for stored procedures etc for actually enforcing business logic (somewhat made possible by Firestore security rules, but very awkward).
This greatly improves latency and simplifies code. Trying to manually merge the mutations into all of your queries would be very awkward.
Also, the whole concept of browser clients interacting directly with the database server is darn interesting. Combine that with better support for stored procedures etc for actually enforcing business logic (somewhat made possible by Firestore security rules, but very awkward).