[09:59:20] meal 2 break [10:00:48] lunvh [10:00:50] lunch [10:24:09] lunch [15:01:57] ebernhardson, dcausse: triage? https://meet.google.com/qho-jyqp-qos [15:02:02] oops [15:56:35] dinner [17:34:22] dinner [19:21:02] this may be a dumb question, but if we can geolocate commons users, how possible is it to only turn on query completion for users in countries/places where the/a primary language is English? [19:24:01] mpham: hmm, maybe? I haven't used geolocation in frontend code but it ought to be possible. The geolocation we have available now is a step performed by analytics as part of ingestion to hadoop iiuc [19:25:02] Looks like geoip is set into the users cookies, so should be possible [19:25:31] that frontend geocoding is not as fine grained as the stuff in hadoop, but should be sufficient i think [19:26:11] alright excellent, should be possible then [19:26:23] I'm wondering if we can potentially sidestep non-English queries getting worse by only enabling for users who are likely using english [19:26:58] though I think it's still a good idea to have ways of measuring non-English queries in place [19:27:15] but it may be a way of not being blocked if they're not great [19:29:28] seems plausible at least. I can't think of anything too crazy involved [19:30:22] thanks for the help! i'll think about it and talk it over with carly and guillaume [20:01:21] I'm talking with Marshall now about T285817. Is there time during hte Wed meeting this week for him, Gergo and maybe one person from Platform to join to discuss what needs to be done? [20:01:21] T285817: Add an image: load static file to search index - https://phabricator.wikimedia.org/T285817 [20:01:36] mpham: should be [20:01:48] thanks. i'll let him know [20:05:54] what were the major questions we had for Growth again? just so i can make sure marshall can prep anything he needs to on his end [20:21:39] hmm, it's not super clear. There is some question of how much the keyword needs to be adjusted to handle their use cases, maybe a bit of a more general thing about how the data is structured. At the end of the day though, is mostly them living with the structure they choose (but we have to support it) [20:24:09] I suppose in part it seems the use case has expanded recently, or perhaps we didn't understand well enough what those use cases are. But it's clear they are trying to adjust the conventions they are using, and we should consider how it best fits i suppose [20:28:07] got it. thanks