[01:11:58] hi [01:14:47] leila: is that possible to translate *.wikversity pages to beta.wikiversity? [01:15:38] My goal would be to translate to esperanto, I saw that MT work prety well for this target language [01:17:20] In fact, my goal is to have an eo.wikiversity.org going on, so I can shamely host my research projects that I intend to write in esperanto there. [15:03:31] _o/ [15:59:32] halfak: found the code I msged you about yesterday, email coming your way in a minute [15:59:44] \o/ thanks dude. [16:01:08] Hi [16:08:41] hiya [16:10:20] Hi Nettrom! [16:46:17] hi hare ! [16:47:29] How is suggestion robot? [16:48:16] it's suggesting things, but probably not in the way you want it to? ;) [16:48:41] * Nettrom needs to get his schedule in order [16:58:06] hi Nettrom [16:58:37] * DarTar reminds you to take a look at your NDA extension (per email) :) [17:00:12] DarTar: replying is on my todo list for today, thanks for the reminder :) [17:30:25] Nettrom: awesome [17:42:20] Nettrom: Can I make a Phabricator task "Implement SuggestBot suggestions for WikiProjects" and assign it to you? I want it to be named explicitly as a blocker to another task of ours. [17:45:45] hare: yeah, please do, having a tracker sounds like good motivation for me to keep moving on this [17:47:42] also Nettrom, if your work has made it to this stage: how does the script determine that a page is within the wikiproject's scope? or does it not care? [17:51:14] hare: currently it does not care, we grab the 128 "oldest" articles from the project's category and find things that are similar to those [17:51:37] Okay. [17:51:41] SuggestBot is very task-oriented, so it'll look for articles tagged with specific tasks, and if it can't find similar ones, it'll just pick random ones, actually [17:52:10] being able to keep suggestions within a given WikiProject is on the wish-list, but will require quite a bit of programming [17:52:12] And you saw in wikiproject.json that there is a source category for where suggestbot=true? [17:52:17] yep [17:52:22] Excellent. [17:52:23] it uses that category [17:52:40] Because WikiProjects do not consistently categorize their shit! I had to write an entire script to generate a database to get around this. [17:54:18] sorry to hear that... those kinds of inconsistencies are rather annoying and tiring [17:55:03] it's one of the reasons why I don't really know how many articles in Wikipedia have quality assessments, I ignored the inconsistent category names [17:55:48] You can get a somewhat precise estimate by going to the WikiProject Directory and seeing what projects have article counts. [18:13:32] o/ hare [18:13:40] Ahoy hoy [18:13:41] Just saw you file the priority prediction one [18:13:46] *one = task [18:14:04] Correct. [18:14:08] WikiProject X is now on Phabricator! [18:14:42] Make sure to use the 'revscoring' tag when relevant :) [18:15:13] The Reverend Scoring has been tagged. [18:15:35] https://en.wikipedia.org/wiki/The_Reverend [18:15:39] New mascot [18:16:00] And new classifier righteous vs. not-righteous [18:55:01] halfak: preferred Phab tag for projects under https://www.mediawiki.org/wiki/Wikimedia_Research/Projects#Contributor_value_research [18:55:24] ideally short and descriptive [18:57:42] the-wikiprice-is-right [18:58:00] I’ll take that :) [18:59:45] My ability to come up with silly project names is only rivalled by Ironholds's. [19:00:33] head to head [19:00:38] :p [19:05:43] DarTar, "value-adding' [19:09:44] halfak: how about value-added-measurement? I feel “value added” might be misread as a qualifier of the task (“by working on this task, we’ll be adding value” ;) ) [19:10:28] Really? That's a lot of characters. [19:10:41] heh I know [19:10:48] va-meter [19:10:54] by working on this task, we will actually not be adding value [19:11:04] "value-adding" != "value-added" [19:11:39] value-measurement ? [19:11:49] I don’t know, you tell me :) [19:12:32] DarTar, sure. 'value-measurement' is fine. [19:12:39] cool [19:14:18] guillom I dont rival? [19:14:58] ToAruShiroiNeko: I haven't had the opportunity to witness your skills in that area. Yet. [19:16:07] guillom really? [19:16:08] oh well, Phab took the project name (Contributor value research) to autogenerate the tag [19:16:16] Everyone knows of Counter Vandalism Unit [19:16:23] That was my creation :p [19:16:38] Oh! [19:16:52] Can you have a sillier name than that? :p [19:17:19] Is that a challenge? :) [19:17:58] It is a challenge I cannot loose [19:18:01] *lose [19:18:35] Reverty Team [19:18:57] * halfak drops mic [19:19:18] heh [19:19:42] * guillom does like the buffyesque -y suffix. [19:58:03] Not feeling great. I'm going to go lay down for a bit. Ping me on gtalk if you need something. [20:20:34] halfak: sorry to hear that, hope you feel better! [21:31:55] DarTar, what's the context on the email you just sent? [21:32:06] (I'm hoping it's a super-exciting collab opportunity but dunno ;p) [21:32:24] Ironholds: I am just about to send you a previous mail as context :) [21:32:28] sweet! [21:32:37] let me find it (I should have done it the other way around) [21:40:28] DarTar, hi! [21:40:56] can I ask you yet another question about EL schemas? :] [21:41:01] sure [21:41:33] I have this PrefUpdate schema: https://meta.wikimedia.org/wiki/Schema:PrefUpdate [21:41:42] yes [21:42:19] it was originally owned by Roan (or maybe you pointed at him when first asked), but he said its a generic thing, not related to flow [21:42:36] yes, it was actually rolled out for Echo notifications [21:42:42] you created it, so, do you know who can own this? [21:42:55] the idea being that the pref table doesn’t have a history [21:43:40] this acts as a log that (1) allows people to understand pref changes (2) map them against the local defaults (which is hard to extract) [21:44:04] aha [21:44:04] that’s a good question, this is a generic log that will serve multiple teams [21:44:04] not tied to any single feature [21:44:16] can Analytics own generic logs? :) [21:44:31] :] I guess so [21:45:12] and, do you know who may be using/interested-in this schema? [21:45:15] that would be my recommendation, it’s important data, it’s going to be hard to find a single owner in Product [21:45:27] ok, makes sense [21:45:36] mforns, 'all of design' [21:45:46] aha [21:45:50] anyone rolling out new features associated with user settings and interested in seeing if people opt in or opt out [21:45:54] seriously you have no idea how many designers I've had sad in the last couple years by explaining we didn't have this in a format that covered all time [21:45:55] yeah [21:46:15] it's a really really fantastic schema to have [21:46:20] cool [21:46:21] agreed [21:46:28] can I make a possibly dumb suggestion? [21:46:33] sure [21:46:40] you might wanna try actually reporting on it. Or heck, I can! Or someone! [21:46:55] like, I imagine a dashboard of "here's the trend in skin usage" would be pretty great to have [21:47:01] we used to have some glorious dashboards for pref changes [21:47:02] I see [21:47:05] ditto conversion-to-VE or something [21:47:14] but that was tied to Echo’s immediate needs [21:47:17] yes, good idea [21:47:26] agreed, this would be really cool data to visualize in aggregate [21:47:46] file the idea on phab? ;) [21:48:13] I don't actually have my phab login today :(. I'm in a log cabin in vermont with 300 pounds of dog and my new (only somewhat password-provided) laptop :( [21:48:57] and I will take advantage of you guys being here: so those pref changes may contain sensitive data? like content related? or they are just data like visualeditor/wikitext, skinA/skinB, etc.? [21:50:47] * Ironholds thinks [21:51:28] the only sensitive data I can think of is say, signature (only sensitive in the sense that it gives you someone else's prefs semi-uniquely if combined with other data) and gender, which is public anyway [21:51:36] the divide is more between 'useful' and 'not useful' [21:52:46] I see [21:55:06] OK, thank you DarTar and Ironholds! [21:56:27] np! [21:56:50] Ironholds: wow, what do you do with a laptop in Vermont? [21:57:07] puts it on the top of his lap [21:58:39] DarTar, try to avoid 300lbs of dog drooling it [21:58:49] also visit a ton of breweries and sit in the sun and go to the range a bit [21:58:50] I actually think we should assume that prefs are private until we’ve done some more serious audit [21:59:13] yeah, makes sense. to be totally clear by 'public' I mean 'if super-aggregated' [22:00:05] DarTar, I've looked at EL PrefUpdate table, and in contains changes on 1919 different settings... [22:00:30] I also think it is good to assume they are private [22:00:38] for example, “timecorrection” will typically reveal someone’s coarse location [22:00:53] I could work on that [22:00:54] Ironholds: yup [22:00:58] I mean, I have their precise location [22:01:01] want me to compare the two? ;p [22:01:08] no, ha ha [22:01:11] I was just saying [22:01:16] "how do you know this isn't geolocating us?" "oh we geolocated you really accura- hey why the angry faces" [22:01:27] I don’t think a lot of people actually change this setting [22:01:47] but Ive never checked [22:02:00] * YuviPanda geolocates Ironholds to a Vermont with a big dog [22:02:35] YuviPanda, you can only go as precise as vermont? [22:02:38] * Ironholds pats yuvi [22:02:47] timecorrection appears 75472 times in the prefupdate event table, and this is higher than most other preferences [22:02:52] (it's two big dogs, I think a 300lb dog would be dead) [22:03:10] mforns: interesting [22:07:30] DarTar, Ironholds, now the million-dollar question: considering the schema can potentially be used to retrieve sensitive data, would you be ok with deleting the userId for events older than 90 days? [22:08:03] I mean, in the data releases you shouldn't have userID at all [22:08:13] I was more imagining you'd just have date-variable-value [22:08:25] we’re talking about internal retention, I guess [22:08:25] aha [22:08:27] 2015-06-07 - "Monobook to Vector transitions" - 12 [22:08:39] for internal retention...hmn. that's a hard one. [22:08:40] not public/aggregate data releases [22:08:53] It would be safest, sure. I can think of use cases for still having the data around [22:08:57] we have a check-in scheduled with Michelle on another log with potentially sensitive info [22:08:57] but not tremendously convincing ones [22:09:00] oh yes, internal retention [22:09:08] I can add this case to the discussion [22:09:11] (e.g. 'get me anyone who had the VE on but turned it off') [22:09:17] indeed [22:10:05] this is really close to a system log, it’s just an accident that it doesn’t exist or that the up table doesn’t retain history [22:10:24] aha [22:10:49] yep :( [22:10:52] see rule 18 [22:11:31] 'a system that is optimised for production is optimised away from research, and vice versa. 18a: nobody builds a successful product that doesn't work in production" [22:12:24] DarTar: timezone preferences don't tell anything more than the edits [22:13:14] Nemo_bis, how? people edit at whatever time they want ;p [22:13:15] hey Nemo_bis: can you expand? [22:14:05] DarTar: did you already forget all the mess ze germans made about the editcounter? :) [22:14:33] well, what I mean is that this is private info [22:14:44] what can be inferred from public data is a separate story [22:14:58] my TZ pref is set to “ZoneInfo|-420|America/Los_Angeles” [22:15:27] FWIW preferences were semi-public in Toolserver for a long time [22:15:28] and I’ve edited from at least 5 different countries in 3 different TZs since I started on WIkipedia [22:15:57] yep, and WMF Legal team was not thrilled to discover that ;) [22:16:17] yet ze germans only found a threat in the editcounter aggregating public info ;) [22:16:51] DarTar: hai poi risposto a quel sedicente giornalista? [22:17:08] DarTar: ho dovuto bacchettare il presidente perché diffondeva dicerie non verificate sulle statistiche ;) [22:17:34] sì gli ho mandato una lunga risposta, mi ha anche chiesto permesso per un virgolettato, ma i miei mi han detto che non è apparso nulla sul Venerdì di oggi [22:17:41] stavo giusto per mandargli due righe [22:17:54] ah va bene [22:18:01] inoltraci la risposta, cosí ci allineiamo [22:18:06] te la giro [22:18:28] ....grappa. [22:18:33] that's all I know [22:19:31] Ironholds: can bring you quite far [22:19:38] you know your basic Italian [22:19:38] DarTar: grazie, ora a nanna [22:19:46] ciao [22:19:57] I can say thank you, you're welcome, goodbye and demand alcohol [22:20:07] what else has a British tourist ever needed except socks and sandals ;) [22:22:46] hey folks, thanks for your help! I'm signing off for today. [22:22:50] have a nice weekend! [22:23:15] have a good one [22:23:39] take care!