Hi, I know that there’s an issue in the rethinkDB git repo that needs to be taken care of before Meteor can support it.
But do you have an idea if this is taken care of ?
I see RethinkDB as a great solution for the realtime scaling problem without changing the api of Meteor (by keeping the same way we use it right now and not using the future data layer based on graphQL and the invalidation server).
Even if it will not be the perfect solution (the db will take the load instead of the meteor server), it’s still way better than oplog tailing and will satisfy almost all the realtime scaling issues that people have.
I wouldn’t expect to see any changes to the database layer until the new data layer project comes to fruition (many months away). At that point theoretically any database could be plumbed into Meteor’s Reactive GraphQL layer, but I think MDG will be looking to the community to make that happen - i.e. to write a reactive handler for the database that meets the Reactive GraphQL standard.
My guess would be that if MDG were going to take on wiring up any other database to RQL it would be after the above is done and dusted, and I’d secondly guess that if they did take that on (which I think is unlikely) then they would go for a straight SQL database as it’s still the item with the most votes on the Trello roadmap, and importantly is the one that potentially opens Meteor up to a larger corporate and SME market.
Just to add - it sounds like the blocker for the meteor-rethinkdb project will be wrapped up by Rethink in the next release (2.3). I agree though - I don’t think there will be much progress on this approach now that Apollo has launched (by MDG anyways).
Same ! Didn’t know about this ! I was going to ask if they would create a simple Meteor client that would be pluggable anywhere, this is the answer haha
More details coming as soon as I can provide them, in the meanwhile we are investigating Relay as a fast-track to getting a Meteor-like experience with GraphQL. Along these lines:
Dude, you the man !! Looking at the picture it definitely simplifies, if it is possible to have a one to one mapping to the existing Meteor experience.
Looking at the diagram Relay and Meteor seems similar at high level. The only thing missing is the optimistic updates on Relay. So here’s the dumb question why can’t tracker be extended as a module to include optimistic updates and caching on Relay as well?
I also think there is some merit in exploring RethinkDb for real time updates? I am just getting in to Meteor ecosystem , just throwing out there to see if it makes any sense in the first place…or if if makes any sense in the first place
