Would it make sense to define a "wire protocol" to make it possible to interface with non-elixir datasets? about loom

[edescourtis]

+1 that

[asonge]

I'm not sure what that would look like. I've been thinking about this and the common use cases for client-side CRDT's and how to evaluate the security implications of merging values from users with different permissions. The only thing I can come up with is that clients that aren't trusted need to use ops-based CRDT's. Doing this in a state-based way seems a lot more difficult and prone to bugs. I'd only use this state-based stuff for maintaining trusted/server state and syncing servers.

When it comes to a javascript companion library that covers most use-cases (and defines a set of json datastructures), it's definitely on my to-do list. My day job should start slowing down soon, so this repo is about to get active again. I think once you have that, it should be possible to write stuff in other languages that obeys the basic json contract.

As far as a protocol goes, this stuff can be made fairly rigorous because actions are idempotent. A complicated protocol doesn't seem to be needed.

[jeregrine]

Just like defining and sticking to a protocol even if its simple can go a long way.

On Wed, Jun 17, 2015 at 5:03 PM, Alexander Songe [email protected]
wrote:

I'm not sure what that would look like. I've been thinking about this and the common use cases for client-side CRDT's and how to evaluate the security implications of merging values from users with different permissions. The only thing I can come up with is that clients that aren't trusted need to use ops-based CRDT's. Doing this in a state-based way seems a lot more difficult and prone to bugs. I'd only use this state-based stuff for maintaining trusted/server state and syncing servers.
When it comes to a javascript companion library that covers most use-cases (and defines a set of json datastructures), it's definitely on my to-do list. My day job should start slowing down soon, so this repo is about to get active again. I think once you have that, it should be possible to write stuff in other languages that obeys the basic json contract.

As far as a protocol goes, this stuff can be made fairly rigorous because actions are idempotent. A complicated protocol doesn't seem to be needed.

Reply to this email directly or view it on GitHub:
#1 (comment)