September sounds good, yeah.

Should we decline this task now that @Smalyshev is using the wiki to test search integration? Stas, how long do you think you will need the wiki?

For this cycle, probably no more than a week. But I think in general lexemes are complex enough (you need some languages, grammatical features, etc.) to bootstrap that keeping around some system which has pre-filled data to play with various stuff for tests might be a good idea. If it's burdensome, I probably could set up one of my own, but I think it may be good to keep this around, since it's already there.

Still needed now that Lexemes are fully enabled on test, beta and production?

Well, if we want to experiment with something and need to do any server-side changes, neither of those are appropriate environments. So this is probably still useful from time to time.

I think we should keep it around for now, unless it becomes maintenance burden. Some things can be tested on beta, but experiments that involve direct-patching configs and code are hard to test on beta, and easy on labs. Main thing for lexemes is that we need a bunch of entities to be defined to make it work smoothly - so having a system sitting around that already has them is beneficial. Absent argument for shutting it down I think we should keep it.

Main thing for lexemes is that we need a bunch of entities to be defined to make it work smoothly - so having a system sitting around that already has them is beneficial.

We also have this need while testing WBQC things.
Perhaps we could create a script in Lexeme to create a base set of items and lexemes for use in testing?

Perhaps we could create a script in Lexeme to create a base set of items and lexemes for use in testing?

That is a great idea, would also allow people like me to set up my own test environment without bothering anybody else.