[05:09:52] <_huh> Hey, I was on this page: https://nrm.wikipedia.org/wiki/Special:AllPages/Talk: and I noticed that one of the pages it linked to just gives an error (https://nrm.wikipedia.org/wiki/Discussion:Catégorie:Calendri_du_meis/Générique) [05:09:55] <_huh> What's going on here? [05:10:28] <_huh> Is it still in the database despite being an invalid page name? [12:50:59] Yeah, that happens. [14:22:15] Hi! I have a technical question about redirects [14:22:40] to create a redirect you can start a page with #REDIRECT [14:23:04] I have seen that in languages other than English there are other ways to make a redirect [14:23:40] for example on itwiki you can use #RINVIA (which is, sorf ot, a translation of "redirect") [14:24:00] This alternative words are defined somewhere? [14:27:08] https://github.com/wikimedia/mediawiki/blob/master/languages/messages/MessagesIt.php [14:30:39] Reedy: thanks, that's exactly what I wanted :) [14:31:01] Then https://github.com/wikimedia/mediawiki/blob/master/languages/messages/MessagesEn.php is probably more inclusive [15:23:01] legoktm: if i strip off the angle brackets, you're left with the mediawiki sunflower that Isarra designed (for the socks) [15:23:15] but it's making a political statement about your preferred sunflower form i guess [15:23:40] there's also the abstract sunflower from https://medium.com/freely-sharing-the-sum-of-all-knowledge/refining-logos-of-wikimedia-projects-a-brand-exercise-190ae689f6aa [15:36:56] That is neither abstract nor a sunflower. [15:37:10] Isarra: i tend to agree. [15:37:20] i'm just saying it's another proposal for "what to do about the sunflower" [15:37:38] and inevitably by adorning oneself with one of the proposals (including the status quo) you're probably taking a mildly political position [15:38:24] Aye, mayhap. [15:38:45] Really I just need to get around to doing an actual official proposal. [15:38:52] Getting people onboard, getting actual feedback. [15:39:22] Otherwise I fear another that is genuinely bad will get declared, and without any proper consensus. [15:40:01] Especially these days where the direction seems to be going toward swapping out logos for iconography. These are different things - icons are generic representations, whereas logos are very specific... [15:40:36] https://medium.com/@cscottnet/i-like-many-of-these-ideas-9c94942b8f <- my thoughts on nirzar's proposals [15:41:32] i'd love to see a guided community-oriented process: here are the elements we've identified as part of the WM "brand" that we'd like to see reused in projects, here are some sample ideas, now you go off and deliberate as a community about what you'd like to do. [15:41:38] Heh, you called them icons. [15:41:46] Aye... [15:42:17] after all, that's how the original logos got built, just over a long period of time and with minimal guidance ("here are three colors") [15:42:23] The problem is, getting everyone involved guiding them through a proper process, that requires time, resources, and especially a particular skill the foundation in general has shown itself to lack in general, let alone with design specifically. [15:43:10] any redesign process that happens in a shorter period of time will probably result in greater artistic consistency naturally [15:43:13] Which isn't to say it wouldn't be impossible to just hire someone specifically for that, but part of the problem seems to be that since we already lack people who do this to begin with, how the hell do we identify someone who can do it well? [15:43:49] It doesn't necessarily need to be shorter, just... specific. [15:44:09] Specific needs and desires and targets specified, leading to specific constraints for all of them... [15:44:48] You write a good spec for what is one of our logos, and everyone actually agrees on it to begin with, I'd argue pretty much anyone could make another down the road that would match just fine. [15:45:21] we do seem to have trouble with bridging the gap between professional designers and community processes [15:47:09] I could be somewhat impolite and take that step further... [15:47:14] perhaps because we implicitly still believe the auteur theory that says all art is the product of one individual's singular and wholly-intrinsic vision [15:47:15] But yes. [15:47:29] Well, that theory is wrong. [15:47:32] Very, very wrong. [15:47:50] yes. but still naively (or just subconsciously) believed [15:48:14] Otherwise there'd be no such thing as principles of design classes, colour theory, accessibility standards, ratio-based aesthetics... [15:48:30] Hm. [15:49:06] anyway, i thought https://meta.wikimedia.org/wiki/Brand/Marks#Making_your_own was a good step in the right direction [15:49:46] "Political statement" for not using what is just a demonstration of something but clearly not a proposal for anything real? [15:49:54] "one individual's singular and wholly-intrinsic vision" is quite useful to have, though. and if you look at successful things, they usually are driven by that one person. [15:50:19] design by committee rarely works. [15:50:25] although it's somewhat undercut by referencing the highly prescriptive https://foundation.wikimedia.org/wiki/Visual_identity_guidelines at the top [15:51:04] Nemo_bis: i'm just saying that "those who know" will recognize that i'm specifically wearing "Isarra's Sunflower" socks -vs- "Old School Sunflower" socks -vs- "Abstract swirl" socks. [15:51:46] DanielK_WMDE: i both agree and strongly disagree. ;) [15:51:49] Vs the many random other svg sunflowrer attempts. [15:52:08] I also agree and disagree. [15:52:15] DanielK_WMDE: i agree it is always useful to have a single (or small #) of architects/editors/whatever [15:52:16] cscott: hehe. "It's Complicated" [15:53:24] Doing commissions for organisations, I've generally followed a process of general to specific - where we start out with general ideas what I or whoever sketch out, and then focus on more specific ones until we all agree - me, clients, any other stakeholders - on what specifically we want for the final product. [15:53:24] DanielK_WMDE: but disagree in the sense that this is often used as an excuse to have the single architect/editor/whatever not listen to anyone else [15:53:38] Adn only THEN are some variants of that made, and one chosen to do. [15:54:00] So yes, you can absolutely do design by committee - you just need good guidance. [15:54:07] But do it right, and everyone's happy. [15:54:33] in this context, i think nirkar's proposal is most useful as a statement of direction, exactly what you'd like an "architect" to state. but i think it could be profitably turned loose to the individual communities to figure out how to apply that direction to their projects. [15:55:03] i suspect wikidata wouldn't be too happy with nirkar's "box in a c" logo for wikidata, eg. [15:55:26] but i bet they could come up with something consistent with his overall design that would work better, just because they know the project better [15:55:48] Maybe, but as is, it's just one proposal. It doesn't actually give the wider project an opportunity to look at various visual options and see what they even mostly want to start with. [15:55:54] The idea that the existing logos don't follow any pattern is just wrong [15:56:33] Isarra: sure. i'm just saying (as I think you are) that the starting point should be meta-designs, not specific designs. [15:56:38] Most people are very bad at coming up with visualisations of their ideas, or acting on how something seems wrong initially, because it's just not what they do. But they can choose between options when they have them in front of them. [15:57:09] Isarra: so we could have a conversation about different meta-designs, select one (or a small #), then solicit input from the community to actualize those meta-designs. [15:57:12] Well, I'm not even sure the starting point should be designs at all. [15:58:09] Nemo_bis: i think there's a different conversation to be had about the fact that the unifying pattern in the existing logos was color, and then we decided to strip color from the brand identity. [15:58:13] That would be a stage after deciding... I'm not sure. Components, style, what we're willing to sacrifice for... things? Complexity vs flexibility... [15:58:17] Sort of thing. [15:59:01] Nemo_bis: but when you put all the logos together it's pretty clear they are divergent [15:59:08] Do we want our general, movement-wide branding style to be more like this image of an owl, this image of an owl that's much simpler, this image of an owl with stylisation, etc... [15:59:11] not even the fonts are consistent [15:59:15] We're not even using owls. [15:59:48] But hells, we need to start with some really basic stuff. How many fonts? How many colours? How many shapes? Lines or no lines? What kind of variants... [16:00:24] Isarra: there's a bit of fashion involved as well. currently the fashion is for flat and abstract. when the wikipedia puzzle globe was designed, it was in a different fashion era. [16:00:37] so i'd be reluctant to say we're going to redesign everything now for all time [16:01:00] i think we should plan to revisit logos and such every (say) 10 years. [16:01:40] and my theory is still that if we do the revisiting in a single year for all projects, you'll always end up with more consistency than if you design the logos over a longer time period as the projects are initiated. [16:28:58] cscott: Just because something is fashionable doesn't mean it's good, or effective. [16:29:34] no, i'm saying the opposite, that we shouldn't assume anything will be good enough to last [16:29:36] But if we do have a good spec nailed down, with clear base things for each project, it should be possible to make variants for different fashions regardless. [16:30:41] As is, we're not even clear on what concept some of the logos are even supposed to be. Is wikipedia a globe or a puzzle or a w? [16:30:51] sure. [16:31:09] Nail down which of those things it should be in general, and you can tweak it while maintaining the identity for your viewers. [16:31:13] and i think wikidata's logo is supposed to say "machine readable" more or less. but there are lots of different ways to convey that perhaps. [16:31:44] And we need to stick to a particular type of one or people will get confused. [16:31:58] wikipedia is "lots of languages put together and there will always be more we are missing" which is, admittedly, a tricky thing to draw [16:33:26] we'll just have to revisit these all over again when we're all wearing magic leap headsets routinely and logos have to be animated and in 3d (but really low res) [17:15:45] Sometimes you also just have to pick some arbitrary visual thing to MAKE into the canonical representation of a difficult concept... [17:17:06] And cscott? I just tabbed back to your response, mistook it for the original thing at first, and then was very disappointed to see that it... wasn't. Because the specific lines I wound up glancing at were some things I've been hoping to see designers accept for... some time now. >.> [17:17:22] Alas. [17:18:36] out of curiosity: which were the lines? (so i can put them on my resume if i ever decide to become a designer :) [17:20:02] Hah! [17:20:27] whoa.... I can't say I was expecting that https://www.drupal.org/drupalorg/blog/developer-tools-initiative-part-5-gitlab-partnership [17:21:00] In particular the thing about colours, where they are useful and should be used where appropriate, but also not always are. [17:29:38] i'll write "smart about colors" on my resume [17:30:04] right next to "former professional lighting designer" ;) [17:43:38] Snrk. [17:48:55] wow "In the past year 111,783 people have contributed to Drupal in some form on Drupal.org" that's nuts, that's similar to the level of contributors we have on Wikipedia https://about.gitlab.com/2018/08/16/drupal-moves-to-gitlab/#how-can-people-get-involved-in-the-project [17:49:25] how is this relevant to mediawiki? [17:49:53] there's been some discussion of Gerrit in the past [17:49:58] so that's why I was sharing [17:50:25] please ignore if you're not interested [17:50:34] I'm always interested in what other projects are doing. :) [17:51:57] that's a good thing, but your first couple messages largely came across as spam to me [17:52:44] oh sorry, not trying to spam. :) [17:53:39] i do agree that it would be interesting for wikimedia to investigate partnerships like this [17:54:08] i've argued in the past that we should narrow our focus and not waste our limited personnel and funds reimplementing certain wheels [17:54:36] (also gerrit is awful) [17:54:53] it's been argued in the past that ownership of our source code and our developer tools was a core mission of the WMF, which was more true when we were using centralized repo tools like CVS [17:54:59] (i like gerrit, fwiw) [17:55:13] yeah I mean that was the argument, they had a patch-based workflow like ours, and it was... uhh.. not really friendly for new devs, so setting up a PR workflow was going to be a whole lot of work [17:55:45] anyway, the interesting thing (to me) is that gitlab is partnering w/ drupal, so basically committing to supporting whatever dev flow drupal wants (within reason) [17:55:45] though it's really similar to the way gerrit works [17:56:13] the partnership aspect is different from saying "we'll switch to github and use whatever tools they decide to provide" [17:56:45] I can definitely see if gitlab comes through for the wmf to perhaps migrate there. wmf on github seems less likely (I know we have mirrors there right now, but I meant for primary development workspace) [17:56:55] simply due to the fact gitlab is open source and github is not [17:57:14] but... lot of work, so the benefit needs to be sufficiently high as well [17:57:54] the benefit is taking all the devs we currently have maintaining gerrit, our git repos, jenkins, etc, and putting them to work on wiki-specific tasks instead [17:58:45] granted, that's not a huge number of devs. and you'd have to weigh that against the increased difficulty of (say) customizing our dev tools for mediawiki [17:58:55] I really hope some of the features that Drupal has on their git workflow makes it into GitLab proper (like attributing commits to the company/client you work for, etc.) [17:58:59] cscott: my issue with gerrit stems mainly around discoverability. It's really hard to browse tasks for a paritcular repo to see what other work is ongoing with it, or to see which patches are stagnating due to lack of review [18:00:12] the search bar in both cases works, but you have to know the exact filter language they use, and this is not exposed anywhere in the UI. Just clicking around doesn't get you far [18:00:20] I mostly like the stack-of-patches workflow, which i recognize not everyone uses. but that's how i like to code. [18:00:37] the commenting workflow could use more work as well, with comments on older patch sets showing up on newer ones as well (but perhaps muted, or being marked as "done" or whatever) [18:01:20] there are some hard choices there about which line (if any) to associate the older comment with, in the newer revision [18:01:39] finally, the fact that gerrit is in its own little world means that it's harder to attract casual contributors vs being hosted on a large platform with a vibrant community working on other things, perhaps leading to less things being fixed than otherwise (I cannot back this up with data, but it seems sensible on its face) [18:01:59] i definitely agree with you there [18:06:41] Skizzerz: Sadly, a Wikimedia-self-hosted GitLab wouldn't help at all with that one (unless GitLab becomes massively more successful so that everyone recognises how it works intuitively – unlikely – and Wikimedia never over-rides behaviour to be different – frankly, with our community, impossible). [18:07:22] why would it be self-hosted? [18:07:36] gitlab.com isn't free software [18:07:46] I'd argue you should maintain a self-hosted mirror to make migration off of the main service easier [18:08:25] a self-hosted anything defeats the final point I just made [18:10:37] legoktm MIT isn't free? https://github.com/gitlabhq/gitlabhq/blob/master/LICENSE [18:11:58] as long as there are multiple vendors who use git infra, i'd argue that maintaining an emergency "migration off the main service" capability is not vital [18:12:24] basically we should mirror all the git repos internally, just to be sure we have a copy if disaster strikes <3rd party host> [18:12:31] davidwbarratt: gitlab.com runs the EE version, which has some code that isn't free [18:12:41] legoktm oh right, yeah that's true [18:12:42] but if disaster struck, we could start doing dev on github (eg) within a day. [18:12:51] it wouldn't be pretty, but it would be enough to disaster recovery [18:12:57] legoktm sorry I was confused on what you were saying. :) [18:12:58] the version, e.g. salsa.debian.org runs is CE, which is fully free (same with GNOME, etc.) [18:13:23] Skizzerz: Anything run for Wikimedia has to be Wikimedia hosted. It's not about the FLOSS side (though that's important too), it's about privacy. [18:14:15] yes, yes you say that. [18:14:17] Anyway, no-one here (especially myself) have any real input to decide these things, so I'd recommend not thinking too much about it. [18:15:51] at any rate, we've outsourced our dev tools effectively to facebook anyway ;-p [18:20:57] Facebook should stay in the FOSS business and get out of the social networking business. :P [18:21:12] all my favorite libs come from Facebook. :P [18:33:04] hi, is there a way to get a list of all missing images in a mediawiki install? [18:35:04] https://en.wikipedia.org/wiki/Special:WantedFiles [18:35:09] Depends what you mean by "missing" [18:35:50] yes that works [18:45:32] is there a way to bulk import images? [18:45:50] yes, there's a maintenance script for that [18:46:00] Blendify_: https://www.mediawiki.org/wiki/Manual:ImportImages.php [18:47:02] that would need to be down by a sysadmin though right? [18:47:44] Yes, from the server itself [18:48:12] If you want to upload from your computer without server access, you'll need something like pywikibot [19:49:19] how exactly does the list all pages feature work? [19:49:37] what does "Display pages starting at:" mean? [19:50:33] I want to retrieve a list of all subpages [19:51:22] Blendify_: it will only show pages beginning from the page with the title you type in there [19:51:34] Blendify_: for that, Special:PrefixIndex may be more convenient [19:52:07] (the difference is, if you have pages titled 'A', 'B' and 'C', Special:AllPages/B will show 'B' and 'C', while Special:PrefixIndex/B will only show 'B') [19:53:28] how can I get the full list? that list gets truncated [19:53:54] I need this to be able to export these pages [19:55:55] i guess just click "next page" and copypaste each separately? :/ [19:56:54] I guess that works [19:58:12] if you use the API, you can get up to 5000 subpages at once. but if you're exporting from the web interface, it's probably easier to copypaste from Special:PrefixIndex [22:18:31] anomie, tgr|away: I'm not sure if I implemented something wrong in my AuthenticationRequest, but on Special:UnlinkAccounts, all I see is a button that says "Unlink", it doesn't show any of the accounts I have linked...is that intended? [22:19:09] also if I unlink my account, and then visit UnlinkAccounts, it lets me submit the form before throwing an exception for having no auth requests