[01:05:20] leila, sorry, just saw this. (too many channels, it was below the fold!) -- You're correct that it's complicated! I don't have a deep understanding of the nuances and terminology. You're correct that user_properties only captures 3 options (male/female/[default]). IIRC, the distinction is that user_properties_table captures grammatical-gender preferences, not actual sex or gender disctinctions. I vaguely recall a mention that there are [01:05:20] some bits of code that interpret the default as "use gender-neutral terms/translations if available" (or something like that?). This task seems to link to all the other related tasks. https://phabricator.wikimedia.org/T61643 [18:19:49] starting in 11 minutes! [18:20:00] you can direct questions to me today [18:22:49] the youtube stream for today's showcase shows the information underneath the video for the last round of talks [18:23:21] Trajectories of Blocked Community Members: Redemption, Recidivism and Departure By Jonathan Chang, Cornell University and so on [18:24:11] thanks apergos ! i'll see about getting that fixed [18:24:16] thanks! [18:25:20] janna_WMF: ^^ [18:26:46] where are you seeing that link? [18:26:46] https://www.youtube.com/watch?v=i9vvwV5KfW4 [18:27:11] yes - in the description below the video there [18:28:21] Oh, I see. Emerauld [18:37:14] thanks Emerauld ! [18:38:00] No problem! [18:46:54] isaacj, Q: If I were going to try to detect policy weaponization computationally, what kind of markers do you think I could look out for? [18:47:10] gotcha -- good question [18:47:18] you want to ask? [18:47:24] Na. go ahead :) [18:51:24] how would you distinguish between 'weaponization' of policies by mariginalized groups in order to try to reclaim space for them to function (in this case, to edit and participate fully as members ot the Wikipedia community) vs weaponization by the dominant group(s)? [18:51:32] in particular for automated detection [18:51:33] New tagline for the Anti-Harassment team: Breaking Civility into Pieces. [18:51:44] gotcha apergos -- i'll try to get that in [18:51:45] isaacj, see apergos's question ^ [18:51:50] :) [18:54:01] * cwylo scribbles down slogan suggestions [18:54:04] excellent question points to apergos ! [18:54:37] thanks for the answer, the difference in tactics is interesting [18:54:51] https://en.wikipedia.org/wiki/WP:OWN [18:54:54] ^ Relevant [18:57:16] good pointing out of norms, but I think we've been dancing around recognising relationships of power and where it lies :) [18:58:22] +1 ^ [18:59:01] I like "power" as a lens for exploring this sort of stuff. Asking who has power or is empower by policy & policy citation practices has brought us far very quickly. [18:59:07] *empowered [18:59:39] Yeah, it's always a fun exercise to read policy pages and ask yourself, who or what has agency in this sentence? [18:59:56] Somewhat related: I'd like to ask, "How do we empower activist groups who are aligned with Wikipedia's mission?" [19:00:29] I like that there are activist groups and they are building common understanding about negative dynamics and how they manifest. I want to go to the next step and ask them/codesign around what might help. [19:00:31] thanks for the WP:OWN link by the way -- fascinating [19:00:36] :D [19:04:51] halfak: I feel like the Women in Red edit-a-thons are a good example of that. [19:05:17] +1 I'm really interested in these kind of activist groups and what they need to get more power :) [19:05:32] It strikes me how both presentations raise the point that policies aren't rules so much as they are tools [19:06:05] That's one of the reasons we're looking into topic routing. One of the things I have heard is that it's hard for these activists to support new article creators because it's hard to find relevant new article drafts. [19:06:10] (cite: Rosiestep) [19:06:47] I want someone to be able to say: "Here's my topic space. I want to grow this community. Give me relevant edits/creations/newcomers/etc." [19:09:56] halfak: remind me that i need to talk to you :) we're working on analyses for the reader demographic surveys that we were running and are looking for ways of representing article topic for looking to see if there are differences in relative # of page views to a given topic between different demographic groups. right now we're at a mixture of drafttopic and wikidata properties (gender, coordinates, instance-of/subclass-of) [19:11:15] Nice. Yeah, I want to pick that embeddings stuff up again soon. Once we get the quarter in order (a couple weeks maybe) i want to get you, dsaez, and one of my engineers together to start working on a pipeline that will serve these kind of needs. [19:12:19] excellent :) [19:14:33] I have this vision of an Ur embedding that is language independent using dsaez's cross lingual mapping strategies. :D If we could build that pipeline, I think it would be a fundamental tool for us. [19:17:25] very much agree -- i've been referring to your post the other day (https://phabricator.wikimedia.org/T221891#5328369). specifically that to understand coverage gaps we need to understand article quality, demand, and topic [19:17:55] I don't want to ask this question to the presenter but I am curious if something like "It would be great if you would pass me the guacamole" is an imperative statement. [19:18:29] isaacj, yes. I like those three dimensions too. I also see those as closing a loop. [19:19:05] Ahh yes I see you referenced where i argued that now ^_^ [19:20:02] yeah - good simple description of the challenge :) [19:23:16] isaacj, Q: Can you reflect on what you have learned about policy usage and what we heard from Elizabeth? Specifically, I'm wondering if there are clear power dynamics at play in policy citation in deletion discussions? [19:23:44] gotcha - want to take it or i can? [19:23:52] Oh yeah. I guess I can do it :) [19:27:35] * halfak goes back to shoveling food into his mouth