: They are experimental not because they don’t work well, but because we : want the freedom to improve them. Starting to actually use them is also : good for confirming the design choices, otherwise we’ll never get any : actual feedback. Similar to the work to migrate the CLI to using V2 : apis.
I think the key queston (and this genuine question from someone out of the loop) is what kind of automation is in place for testing the UI? Because the advantage that comes from freedom to improve the V2 APIs (w/o heavy consideration of backcompat) becomes a liability if people make big improvements/changes w/o realizing that doing so breaks the UI. if tweaking an API in a way that breaks the UI (or CLI, or SolrJ, etc...) is something that immediately and obviously causes a test failure, then that's far less concerning because if the the "client" can be considered a blocker for that commit. -Hoss http://www.lucidworks.com/
--------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
