[12:54:43] halfak: I'll be at our meeting [12:55:09] an assertion that is sadly necessary due to how many times I missed it :) [13:00:23] :P [14:31:37] Hello. [14:45:51] hi [14:51:35] Good TAG folks :) [14:51:48] TAG == Timezone apt. greeting [16:14:46] Guten tag for you as wel [16:15:55] :) [16:16:04] Gluten tag? [16:16:10] Or guten free tag? [16:17:10] on Fridays you can go "guten freita", I think ;) [16:17:12] o/ [16:17:18] +g [16:17:46] Where to put the "g"? [16:17:58] at the end [16:18:00] Or are you adding group rights ;) [16:18:10] Ahh! [16:18:18] freitag! [16:18:54] that's about as much German as I know [16:19:53] http://www.urbandictionary.com/define.php?term=DU+WERDEST+EINEN+KRANKENSCHWESTER+BRAUCHEN [16:24:26] Hallo hare! [19:23:40] hellooo J-Mo. [19:23:52] question for you: was your team involved in https://docs.google.com/forms/d/1ughvbZgRzVvlTpA2pP5-LkdjzXuWQnRL0aSnLa51OpQ/viewform ? [19:24:28] I just saw it, and I'm wondering if you guys know about it since it definitely has some impact on how you guys run surveys. [19:24:55] I'll respond to the phab task once I hear from you, J-Mo. [19:25:08] leila, no I wasn't involved in this. probably doesn't affect us, since we aren't requesting support from L&E. [19:25:50] well, they're telling us that the Tool Labs survey we want to run should be coordinated, since their team will be running a lot of surveys in August, and they want to make sure there are not burn outs. [19:26:19] I'll add you to the phab discussion about it. [19:26:43] cool [19:46:29] leila: I wonder if that should affect us since I don't know if there are going to be overlap in the users surveyed [19:46:38] I'm in Chicago airport now. Gonna check mail [19:51:36] To be honest, I'd worry less about the number of surveys, and more about the length and the validity of the questions, both of which can cause respondents to rage-quit. [19:52:05] leila: I'm editing minor bits [19:52:34] leila: looks great! slightly worried about page 4s wording, but can't think of anything better atm [19:53:11] there will most probably be overlaps YuviPanda but it's okay! [19:53:54] I tried to respond to the Wikimania survey, and was already getting a bit annoyed by how long it was taking, but I quit when they started asking questions about the session tracks, which were basically made up. [19:54:02] yes, guillom. I'm with you, though the number can also be problematic. Assume a user receiving three surveys from the WMF in one day. It's unlikely, but it's annoying probably [19:54:29] leila: Agreed; I'm not saying it's not a problem. I'm saying that from my experience, it's not the only one :) [19:55:01] I also don't think we should overoptimize this thing. we should have a page that people put the dates they want to run surveys, the population, and the team running it. others should use common sense to avoid running surveys on those same dates unless they really have to. :-) [19:55:42] omg. Wikimania survey! I started it 3 times, the first two times, I gave up in the middle. :D [19:56:04] the last time, I really tried hard and made it to the end, for the good of humanity. ;-) [19:56:13] leila: Similar to the CentralNotice calendar, then. But I think their goal is also to help design surveys, because we (the WMF) have a tendency of doing surveys for many things, and not necessarily well. [19:56:25] heh, I wasn't even trying [19:57:27] guillom: it will be an interesting discussion on Thursday since J-Mo's team will probably want to keep the way they design surveys under their control. That's in their title. :-) [19:57:36] :) [19:57:39] YuviPanda: you should have! [19:58:42] leila: I'm quite happy with what we have now, btw :) do you mind if I involve the other admins? [19:58:42] YuviPanda: wording needs major fixes from a native speaker. [19:58:49] leila: +1 [19:58:52] not at all. please. [19:58:55] do you know how? [19:58:55] :D [19:59:20] File -> Add Collaborators YuviPanda [19:59:27] yup! [20:00:21] and just a note, YuviPanda. Legal has not vetted on Google Form yet (they are working on it and they get back to us before Friday). It shouldn't affect the way we think about the survey, but we may have to move to Qualtrics or somewhere else (which I do if necessary) [20:00:54] leila: yeah, that's fine [20:01:00] guillom, leila I also stopped filling out the Wikimania survey after the first few pages ;) I support a survey calendar. I won't be submitting my drafts for mandatory review, tho. I'm pretty confident in my ability to write survey questions well and scale them appropriately (as well as ask for input when I know my work needs sanity-checking) [20:01:57] J-Mo: I agree; I think people who are not expert at designing surveuys should be uncouraged to seek help, but not that it should be mandatory. [20:02:11] yeah, I didn't know about the existence of such a calender. Having it public helps, too, so others can see what's coming on their way, J-Mo. [20:02:19] sounds good to me. I'm always happy to review other peoples' drafts [20:02:28] yeah, needs to be public for sure. [20:02:46] * YuviPanda adds J-Mo to the toollabs survey draft :P [20:02:59] * J-Mo shuts his big mouth [20:03:01] seriously, YuviPanda. Just thought to do the same thing. :D [20:03:02] * guillom 's first work for the WMF was to do a survey, and it wasn't very well designed. [20:03:13] hahaha :D [20:03:19] we've all been there [20:04:39] leila: I'll add the others as soon as I get their google accounts confirmed :) [20:04:48] J-Mo: do check it out, etc though :) [20:05:04] oooh, rdaiccherlb is here [20:05:09] oooookkkkaaaaaaaaaayyyyy YuviPanda [20:05:15] J-Mo: +1 :) [20:05:41] sounds good YuviPanda. :-) [20:05:44] J-Mo: quarry has taken a backseat to ORES for now :( [20:06:04] * rdaiccherlb freezes. She's been spotted [20:06:11] GET HER! [20:06:23] GHAH [20:06:36] You shan't leave ever. [20:06:40] Meh. [20:06:45] hahahaha [20:06:46] lol [20:07:46] YuviPanda oh well. It's a good product, and it's working. We have a good list of prioritized fixes and feature requests, so we can pick it back up in The Glorious Future In Which We Get To Work On What We Want. [20:07:55] +1 [20:08:06] J-Mo: or when I can offload all of ORES work to madhuvishy :D [20:17:02] YuviPanda, :( [20:17:05] madhuvishy, :D [20:17:14] halfak: :( on? [20:17:16] oh, ORES [20:17:16] heh [20:17:18] :P [20:20:31] YuviPanda, where would one "enable CORS" to allow json requests to our WMFlabs servers from wikipedia.org? [20:21:50] It seems like we would need to do something on the wikipedia.org domain. [20:21:53] halfak: it's a header you can set in Flask or I can set in nginx [20:21:54] indeed [20:21:58] I opened a bug for it a long time ago [20:22:18] Yeah. I'm looking at it now. It seems the bug is filed against ORES. Is that the right place? [20:22:26] halfak: yup, it's a HTTP response header [20:22:36] But why do *we* get to set it in ORES? [20:22:48] shouldn't it be enabled on wikipedia.org? [20:23:41] As in, the authorization to use something on wp.org from labs should be handled where it's actually used? [20:23:49] yeah [20:23:55] Thanks guillom. nicely put [20:24:09] * guillom is The Reworder. [20:24:43] halfak: ah, no. so it's us allowing ORES to be called from JS in wikipedia.org [20:24:51] It seems like the security model would require this. As the server (in labs), I don't care how the request comes. [20:25:02] yes, so ORES doesn't care because it's fully readonly [20:25:09] As a site owner, I want the javascript that runs on my site to not be able to call external services. [20:25:22] but imagine you can send JS 'logout' requests to wikipedia.org from random-phishing-website.org [20:25:34] k [20:25:39] or an edit js request [20:25:40] Why would that be bad/ [20:25:44] ? [20:26:02] because I can steal your cookies and make edits pretending to be you by just tricking you into visiting my harmless looking blog [20:26:24] Hmm... I dunno how you get to read my cookies. [20:26:26] But OK [20:26:31] if you're logged into wikipedia and a request is sent to wikipedia from your browser it takes your authorization cookies with it by default [20:26:48] no matter which site the js originating the request is from [20:26:55] but anyway, ORES is a special case since we totally don't care at all [20:26:58] (same with Quarry) [20:27:28] Huh. [20:27:59] halfak: https://github.com/wikimedia/analytics-quarry-web/blob/master/quarry/web/app.py#L248 [20:28:54] Seems like nginx is the right spot [20:29:05] Given that ORES could run on other domains and stuff. [20:29:33] yeah [20:29:44] I'd like someone not me to make that change, but if it's urgent I can too [20:29:46] it's about 2 lines ong [20:29:47] *long [20:29:51] * YuviPanda looks at madhuvishy :D [20:30:00] I'm totally ok making it if you're bus [20:30:00] * madhuvishy stares back [20:30:00] y [20:30:15] madhuvishy: think you can add a puppet patch to enable CORS for * on ORES? just some nginx config... [20:30:23] YuviPanda: I can make it. [20:30:29] WOot! Thanks. [20:30:36] madhuvishy: w00t! cool. [20:30:41] BTW, I'm trying to figure this out because it is starting to block a lot of things. [20:31:08] oh heh [20:31:23] it's the 'right' way to do jsonp [20:46:59] omg rdaiccherlb is here again [20:47:03] don't let her leave this time!!! [20:51:28] YuviPanda: https://gerrit.wikimedia.org/r/#/c/229212/1/modules/ores/templates/lb.nginx.erb [20:56:24] o/ rdaiccherlb [20:56:25] :) [20:56:43] Research *is* one of the best parties on IRC. [20:56:44] I will probably have a meeting soon [21:00:19] halfak. where are you? ;-) [21:01:01] halfak: CORS should be enabled on ores now [21:01:30] \o/ I'll runs some tests. [21:01:33] Thanks madhuvishy ! [21:02:13] ah [21:02:13] heh [21:02:15] so many channels [21:03:34] madhuvishy, what should I link to when noting the the change has been made? [21:03:38] A patchset? [21:03:43] https://gerrit.wikimedia.org/r/#/c/229212/ [21:03:47] Thanks [23:52:56] Hi everyone! I want to get the list of articles created on a particular day. [23:53:08] Is it stored anywhere? [23:53:51] I found halfak's project that gets this info, but over all the articles: https://github.com/halfak/Wikipedia-article-creation-research/blob/master/sql/enwiki/pages.table.sql [23:55:02] I am running the second half of the query given there, but since it runs through the entire revision table, it has been running for a long time now.. [23:55:36] Any help is great :-)