[00:00:10] <wm-bb>	 <amire80> OK, I'll make a patch and I'll document the ZID in qqq.
[00:00:20] <wm-bb>	 <vrandecic> Thanks
[00:00:53] <wm-bb>	 <amire80> The reason I've arrived at this message is different, and it will become clear from a task I'll file in a couple of minutes ;)
[00:01:40] <wm-bb>	 <vrandecic> I'm surprised this message would ever show up, but I see it's in the API documentation
[00:02:05] <wm-bb>	 <amire80> You know what is really a good surprise?
[00:02:21] <wm-bb>	 <amire80> Not a new surprise at all, but it's one of the most inspiring things.
[00:02:42] <wm-bb>	 <vrandecic> I'd love to hear
[00:03:02] <wm-bb>	 <amire80> The apihelp messages have been translatable for ten years. One could think that it's not necessary: all developers know English anyway, don't they?
[00:03:34] <wm-bb>	 <vrandecic> That has never been my argument :D
[00:04:21] <wm-bb>	 <amire80> However, more than once I've heard from gadget developers in non-English wikis that they are happy that the apihelp is translatable and translated to their language. They use the ApiSandbox and it's more convenient in their languages.
[00:04:37] <wm-bb>	 <amire80> Of course, but it is many people's argument 🤷🏻‍♂️ (re @vrandecic: That has never been my argument :D)
[00:09:02] <wm-bb>	 <amire80> Oh, and "String" is Z6 and "Boo" is Z40, right? (re @vrandecic: It should be typed pair, indeed)
[00:23:41] <wm-bb>	 <amire80> https://phabricator.wikimedia.org/T360588 / https://gerrit.wikimedia.org/r/c/mediawiki/extensions/WikiLambda/+/1013154
[00:28:56] <wm-bb>	 <vrandecic> Yes (re @amire80: Oh, and "String" is Z6 and "Bool" is Z40, right?)
[10:24:16] <wm-bb>	 <waldyrious> That commit message is not totally accurate ;) (re @amire80: https://phabricator.wikimedia.org/T360588 / https://gerrit.wikimedia.org/r/c/mediawiki/extensions/WikiLambda/+/1013154)
[10:25:12] <wm-bb>	 <waldyrious> (though I do see the nice to have bit mentioned in the Phabricator task)
[11:25:28] <wm-bb>	 <amire80> Better now? (re @waldyrious: That commit message is not totally accurate ;))
[11:52:02] <wm-bb>	 <waldyrious> Yes! Thanks for indulging me 😇
[19:17:02] <wm-bb>	 <vrandecic> Newsletter 148: Towards internationalizing numbers
[19:17:03] <wm-bb>	 <vrandecic> 
[19:17:05] <wm-bb>	 <vrandecic> https://www.wikifunctions.org/wiki/Wikifunctions:Status_updates/2024-03-21
[19:18:42] <wm-bb>	 <Al> i18n in many languages and scripts, anyone? (re @vrandecic: Newsletter 148: Towards internationalizing numbers
[19:18:42] <wm-bb>	 <Al> 
[19:18:44] <wm-bb>	 <Al> https://www.wikifunctions.org/wiki/Wikifunctions:Status_updates/2024-03-21)
[19:19:34] <wm-bb>	 <vrandecic> I know that @ebraminio was also particularly interested in this topic, and I wanted to point out we are close!
[19:20:34] <wm-bb>	 <vrandecic> Yes Al ! We need now functions to display and read numbers in many scripts and for many languages
[19:20:58] <wm-bb>	 <vrandecic> And editorial decisions on which numbers to use for a given language setting
[19:21:17] <wm-bb>	 <vrandecic> But we also have some trouble with running the whole thing
[19:22:47] <wm-bb>	 <Al> Ah, yes… lazy orchestration, anybody? (re @vrandecic: But we also have some trouble with running the whole thing)
[19:26:09] <wm-bb>	 <ebraminio> that's fantastic, thanks for letting me know (re @vrandecic: I know that @ebraminio was also particularly interested in this topic, and I wanted to point out we are close!)
[19:27:41] <wm-bb>	 <vrandecic> Al it seems more to be a network issue than evaluator strategy issue, but we really lack measurement data. We are aware and planning towards getting more insights. Until then, I hope some ingenuity can resolve this issue somehow.
[19:32:40] <wm-bb>	 <Al> The Z20 details provide some end-to-end perspective. We do have quite a few of those now. But, very much not my field, I’m afraid. (re @vrandecic: Al it seems more to be a network issue than evaluator strategy issue, but we really lack measurement data. We are aware and plan...)
[19:40:48] <wm-bb>	 <Al> Just thinking aloud, that even when an evaluation needs to be repeated, it will often be the case that the orchestration is the same each time 🤔 (re @vrandecic: Al it seems more to be a network issue than evaluator strategy issue, but we really lack measurement data. We are aware and plan...)
[20:15:35] <wm-bb>	 <vrandecic> We have some form of caching on for those things, but not as much as I'd like
[20:15:42] <wm-bb>	 <vrandecic> that would be one place where we could improve
[20:40:33] <wm-bb>	 <Al> Keep at it! If you need us to try anything out, you know where we are. (I had hoped that Reduce would scale better, given that it’s the same function in each evaluation, but it always seems to fail if there are more than four elements in the list.) (re @vrandecic: that would be one place where we could improve)