[17:59:11] the meeting starts in a minute, yeah? [17:59:21] that’s the schedule i’ve got yes :D [17:59:25] hi, one or two, I will facilitate it [17:59:36] great [18:00:25] what's the meeting about? [18:00:33] Vertical writing support [18:00:40] oh cool [18:00:53] ok, just finished a meeting, landing here [18:01:44] brion, you're here. good. can we start? Tim-away ? [18:01:59] i think we’re good to go [18:02:07] it’s a bit early for tim [18:02:17] (or late) :) [18:02:18] ok [18:02:20] :D [18:02:26] #startmeeting [18:02:26] qgil: Error: A meeting name is required, e.g., '#startmeeting Marketing Committee' [18:02:51] #startmeeting Architecture RfC meeting [18:02:51] Meeting started Wed Jul 16 18:02:51 2014 UTC and is due to finish in 60 minutes. The chair is qgil. Information about MeetBot at http://wiki.debian.org/MeetBot. [18:02:51] Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. [18:02:51] The meeting name has been set to 'architecture_rfc_meeting' [18:03:06] #link https://www.mediawiki.org/wiki/Architecture_meetings/RFC_review_2014-07-16 [18:03:17] So, vertical writing support in Mediawiki. [18:03:21] Alright, this we have one topic to focus [18:03:44] #topic Vertical writing support [18:03:50] #link https://www.mediawiki.org/wiki/Requests_for_comment/Vertical_writing_support [18:04:06] YairRand, all yours [18:04:10] People who stand to benefit from support for vertical writing include the millions of people whose native language is a sign language, Inner Mongolian speakers, and many speakers of Chinese, Japanese, and Korean. [18:04:17] (CSS spec for vertical text is at http://www.w3.org/TR/css3-writing-modes/#vertical-intro ) [18:04:25] Vertical text is supported by IE8+ (with partial support in 5.5-7), Chrome, Safari, and Opera 15+, but is not yet supported by Firefox. (Mozilla Bug: https://bugzilla.mozilla.org/show_bug.cgi?id=145503 ) [18:04:34] (Some of these have moderately buggy support, but there aren't any really serious issues as far as I can tell.) [18:04:43] تحذير [18:04:43] warning [18:04:43] you may be watched [18:04:43] do usa&israel use the internet(facebook,youtube,twitter, chat rooms ..ect)to spy?? [18:04:43] do usa&israel use the internet 2 collect informations,,can we call that spying?? [18:04:43] do they record&analyse everything we do on the internet,,can they harm you using these informations?? [18:04:43] In order to support vertical writing, CSSJanus needs to be extended to be able to support vertical text, and scripts need to be modified to work with the different page direction. [18:05:01] A lot of scripts access or modify direction-related styles which need to be changed in a vertical text environment. For example, width needs to become height and vice versa, top <-> left, bottom <-> right, margin-top <-> margin-left, and so on. [18:05:03] tesh please respect our meeting [18:05:10] YairRand: what’s the behavior in Firefox? is it sorta legible or totally broken? [18:05:17] totally broken [18:05:22] doesn't react to the styles at all [18:05:24] heh ok [18:05:27] qgil: is a spambot, has been flodding other channels recently. kickban would be appropriate [18:05:43] worst case we devise a browser warning for firefox until they get vertical support in :( [18:06:22] Even more difficult to deal with is that many scripts call direction-related functions upstream. Example: (mediawiki.toc.js) clicking on the show/hide button in the table of contents calls jQuery.fn.slideUp/Down on the ToC. There are no horizontal equivalents of those functions in jQuery. [18:06:35] #info Vertical text is supported by IE8+ (with partial support in 5.5-7), Chrome, Safari, and Opera 15+, but is not yet supported by Firefox. (Mozilla Bug: https://bugzilla.mozilla.org/show_bug.cgi?id=145503 ) [18:06:41] not sure what can be done about that [18:06:53] #info upstream uses of directiony things like jquery animations may be tricky too [18:07:19] ah, "partial support" in IE5.5-7 refers to supporting vertical-rl but not vertical-lr [18:07:22] those could potentially be redone with css transitions etc [18:07:55] For those who aren't familiar with how a vertical-writing oriented page would look, please take a look at https://incubator.wikimedia.org/wiki/Wp/ase/AS18517S20500S2ff00M529x544S2ff00482x483S20500519x504S18517503x517 (assuming you have a supporting browser, i.e., not firefox) [18:08:04] Or, if you don't have a supporting browser, https://commons.wikimedia.org/wiki/File:SignWritingIncubator.png [18:08:21] #link https://incubator.wikimedia.org/wiki/Wp/ase/AS18517S20500S2ff00M529x544S2ff00482x483S20500519x504S18517503x517 [18:08:27] things turn sidewards :) [18:08:29] #link https://commons.wikimedia.org/wiki/File:SignWritingIncubator.png [18:08:48] YairRand: is that implemented with some javascript overrides or something? [18:08:54] seems to render in safari :D [18:08:56] brion: yes [18:09:23] js actually is just loading a big manually-rotated version of vector.css [18:09:43] #info current signwriting incubator uses some javascript/css overrides, manually rotated css. need to automate this process into RL [18:10:18] The SignWriting Gadget contains the code used to transform the display. https://incubator.wikimedia.org/wiki/MediaWiki:Gadget-Signwriting.js [18:10:29] #link https://incubator.wikimedia.org/wiki/MediaWiki:Gadget-Signwriting.js [18:10:29] nice [18:10:47] supporting optional vertical writing in Chinese, Japanese and Korean comes with the additional issue of needing to be able to rotate existing wiki page CSS on the spot. [18:11:16] *nod* [18:11:40] I think the easiest to start with would be just getting vertical writing to work with the ASL wikipedia and such [18:11:43] YairRand: do we want to concentrate on the sign languages case or also include the CJK? is it simpler to do the vertical-onlies? [18:11:46] ah perfect :D [18:11:56] did someone actually check with natives if putting page layout on its side is good for them? because monitors are still mostly wider horizontally [18:11:58] i tend to agree [18:13:02] ^ re starting on the sign language [18:13:37] MaxSem: I haven't asked any native speakers regarding rotating the whole skin layout, no. [18:13:50] but I strongly suspect that it's the best solution [18:14:07] I guess it depends on the language and perhaps even content [18:14:10] otherwise tricky to do UI elements i guess [18:14:24] quite [18:14:29] MaxSem: I thnk you have a point there, especially with wide screen, but the correct writing rotation is still what a native speaker would expect IMO. Some modern screens can be rotated, so that would solve that issue anyway. [18:15:32] shuld we also consider doing article content and interface support as separate steps? [18:15:41] Regardin the screenshot, the stuff on the left is from the bottom up rather than top-down. Is that by design or WiP? I'm not a native BTW, just thinking out loud [18:16:03] Nikerabbit: hmm, that might make sense for CJK but not sure how that would work for ASL/signwriting [18:16:21] GunChleoc: the left goes to the top and vice-versa [18:16:40] also, gotta have a way to REALLY QUICKLY switch the layout to the traditional one, otherwise visitors from other wikis will be lost [18:16:44] since the search box is on the right, it gets moved to the bottom [18:16:58] even finding prefs is hard [18:17:07] brion: I guess CSS Janus work is needed for both, and that is a big chunk? [18:17:28] MaxSem, "Back"? [18:17:47] OK I just thought that it looked weird to have the writing top-down and the menus bottom-up, but it that's how it's supposed to be, OK. Just checking ;) [18:18:15] how's the progress on getting signwriting into unicode? [18:18:24] hmm, on macs mouse wheels aren't scrolling horizontally-oriented pages [18:18:29] GunChleoc: the menus are top-down as well; the ones at the bottom are the ones that are right-aligned in english :) [18:18:41] Nikerabbit: getting it into unicode isn't strictly necessary [18:18:45] MaxSem: use trackpad like steve jobs intended [18:18:59] steve jobs is dead [18:19:08] slevinski, care to explain the current method of getting around unicode? [18:19:13] YairRand: I agree, but still something I would aim for [18:19:13] and my hand is hurt from using is heavily [18:19:30] SignWriting in Unicode is a work in progress [18:19:31] Oh, I get it now. Shouldn't the lillle search magnifying glass be at the top of the field though, since it's in the position where one starts typing? [18:19:53] [ the only website I know that actually uses vertical writing is the site of the Communist Party of China: http://mongol.people.com.cn/ ] [18:19:57] GunChleoc: in horizontal writing, is the glass at the start or the end? [18:19:59] :) [18:20:07] works quite nicely in Chrome [18:20:08] The SignWriting symbols themselves are proposed, but not the 2-dimensional layout. [18:20:41] The current technique for SignWriting uses a lite ASCII markup. The page link above contained some of that code, called Formal SignWriting [18:20:46] YairRand: Duh, I never noticed that you stuck the thing at the end [18:20:52] I'll shut up now *lol [18:21:00] also, with current layout pages are oriented from left to right but english chars are displayed with right side as their "top" [18:21:09] ie kinda upside down:) [18:21:21] figuring where things are supposed to be in these layouts is... unintuitive :) [18:21:33] MaxSem: it looks a little odd, but is consistent with embedding english in CJK vertical text [18:21:41] The ASCII markup is given a zero size font and the SignWriting image is pulled in as a background image [18:21:44] keeps the direction flowing right [18:21:44] "little" :P [18:21:47] Just turn your head to the left and imagine wou're writing from right to left, is that it? [18:22:01] GunChleoc: that'll work [18:22:39] The SignWriting images are created by a SignWriting Icon Server running on Wikimedia Labs. [18:22:55] I guess what got me dorn the gearden path with the search box is the missing "Search" title, which would be displayed top to bottom [18:23:11] We have an experimental TrueType font that shows promise, but it is not production ready yet. [18:23:24] ah, certain browsers don't handle input boxes very well in vertical orientation [18:24:01] as in, it's still sort of acting like a horizontal element [18:24:19] thus, no large visible "Search" going from top to bottom [18:24:43] Sounds like you got your work cut out for you [18:24:44] ASL WP is probably going to need a custom solution to input boxes anyway, though [18:25:16] I believe slevinski is working on a special input for signwriting, yeah? [18:25:21] *input method [18:26:05] Yes, the special input for SignWriting is a work in progress [18:26:17] anyhow, major issues from the mediawiki side: rotating scripts (probably rather difficult) [18:26:50] there are also minor things like that image thumbs need to assign height instead of width attributes [18:27:03] (currently being dealt with using js on incubator) [18:27:58] though honestly, I'd consider only the lack of CSSJanus support to be a completely blocking bug regarding the ASL WP [18:28:25] everything else is pretty much inconveniences [18:28:38] (on mediawiki's side) [18:28:47] YairRand: so who’s going to work on the CSSJanus tweaks? [18:28:56] you guys ready to do that work or you need help with it? [18:29:18] I personally do not actually have the relevant skills for that [18:29:34] I'm bit wary when I see "custom solutions" mentioned so often. We need to ensure all relevant parts of code can handle this like they do for RTL currently [18:29:48] #action group working on ASL needs someone to help with the CSSJanus improvements [18:30:15] I hear from theDJ's summary of the arch summit that "Roan & Trevor will gladly help anyone to do this" [18:31:29] one other minor issue is how to display vertical languages in interwiki boxes [18:31:36] they don't quite fit [18:31:42] :) [18:31:57] of course, sign languages in particular have a larger issue of how to get the characters to display in the first place [18:31:58] YairRand: is there a semi-standard way of squishing them into horizontal view? (eg standard rotation) [18:32:06] worst case: SVG ;) [18:32:12] I think yes [18:32:18] not positive [18:32:22] slevinski? [18:32:47] Horizontal is possible [18:33:07] in sign languages in general? [18:33:15] Some information may be lost, such as body weight shifts to the left and right [18:33:32] but that generally wouldn't be relevant in a single language's name, right? [18:33:52] Not for a name [18:33:59] excellent [18:34:23] of course, getting the image in there is still potentially an issue. [18:34:28] #info need to work out lang name display also [18:35:09] Another issue: Certain images need to be rotated. [18:35:25] Is this handled by CSSJanus for ltr/rtl, btw? [18:35:35] The vertical layout uses lanes (left, middle, and right) for discussions. Horizontal writing would loose the lane information as it does not use lanes. http://signpuddle.net/wiki/index.php/MSW:Script_Layout#7.C._Vertical_Layout [18:35:57] Only punctuation would need to be rotated to switch from vertical to horizontal. [18:36:25] I meant the mediawiki icons, actually [18:36:35] arrows and such [18:36:44] YairRand: so there’s ability to specify rtl overrides for icons which the flipping can use automatically [18:36:58] we could probably devise a vertical extension to that yeah [18:37:07] #info make sure icon flipping is extended to vertical as well (arrows etc) [18:37:23] the override at incubator is also using manually rotated images [18:37:34] not a good long-term solution :) [18:38:01] yeah the other possibility is to apply CSS transformations but that way may lie madness ;) [18:38:24] can those even be applied to background images? [18:38:37] mmm, you can transform an entire element, that owuld include the bg i think [18:38:52] but i’m not sure you can transform a background image *without* transforming the contents [18:38:54] of the delemnt [18:38:56] *element [18:39:18] wouldn't work for things like the vector tabs shading [18:39:35] hmm those should really be CSS gradients anyway :) [18:39:51] or the little arrow in the "More" box, I think [18:40:17] actually maybe it could [18:40:29] I dunno. probably not going to be the method used anyway. [18:41:22] (Hm, I feel like I missed one of the issues...) [18:41:28] YairRand: so going forward — you are the main contact on this? [18:41:39] apparently :) [18:41:49] great :D [18:42:16] YairRand: do you want to check in with roan or trevor about cssjanus? [18:42:33] and if they don’t have time to poke at it right now, i’ll see if i can help [18:43:13] I don't have any useful skills to add on that front, which makes me question the whole me being the main contact on this thing... [18:43:27] :) [18:43:37] anyone else able/willing to take on part of that? [18:45:46] silence :) [18:45:53] I am curious of this project, but I am not Janus expert either [18:46:00] :D [18:46:13] YairRand: ok sounds like you’re still point person but you need a technical liason :) [18:46:20] :) [18:46:48] ah, other point: (copied from theDJ's summary) "we need a language -> rotation map, like we have for rtl/ltr directionality. Ideally we would get this information from CLDR (http://unicode.org/cldr/trac/browser/trunk/common/properties/scriptMetadata.txt), but it is not in that so we need to bolt it on top of our language data first. (the Translate/Localization team can be involved)" [18:47:41] I could do some small tasks and help where I can, if you come up suitable tasks, like the one above ;) [18:47:49] awesome [18:47:57] wonder if we can upstream that to CLDR also [18:48:14] #info CLDR missing language->verticalness mappings, can we upstream these? [18:48:23] brion: prolly, can talk to Amir & Nemo_bis when we know what we want [18:48:45] I already have a list of sign languages in ISO. there aren't all that many vertical languages (I think), so it might be simple enough to just build a list manually. [18:49:00] (I hope, at least) [18:49:11] Maybe you can poke the people at CLDR? They areworking on a new version at the moment [18:49:37] I guess it's a bit chicken and egg, but if we already use certain info in MediaWiki, that could convince and speed up CLDR processing [18:49:53] hm, perhaps [18:49:54] *nod* let’s probably start with a local mapping and then try to push it upstream [18:50:22] (10 minutes to go) [18:51:03] ok i’m willing to look into the cssjanus provisionally, can take some of my research time for that [18:51:07] Give them the info you have when you poke them, that might speed things up. I've only been ilvolved with them as a translaor so far, so I don't know how the techy stuff works. They do have a bug tracker though [18:51:10] and will escalate to roan/trevor as necessary :D [18:51:25] excellent :) [18:51:36] YairRand: anything else we need to plan before wrapping up for now? [18:51:42] next checkin date for the rfc? [18:52:00] #action brion to look into the cssjanus provisionally [18:52:01] YairRand: oh — anyone working on this going to be at wikimania? [18:52:10] we could do some testing there [18:52:40] brion: I'm not going to be at wikimania. [18:52:51] dunno regarding anyone else [18:53:12] ok then we’ll probably plan to regroup online sometime after [18:53:32] qgil: any recommendations on scheduling, or leave that to later? [18:54:03] It would be good to have a tentative date that is realistic, or some measure to make sure the discussion and work doesn't die... [18:54:27] YairRand seems to be quite dedicated, though. :) [18:54:50] :D [18:55:07] YairRand, will you want to summarize the status after Wikimania, and find then a tentative date for the continuation of this RfC? [18:55:27] qgil: sure [18:55:50] I'm not sure how I'm going to find out what happens at wikimania, though [18:55:55] i’ll probably initiall test with japanese as i am slightly better at reading that than signwriting :) but i’ll just start with the flipping code and that should apply to both [18:56:09] YairRand: i’ll send mails to wikitech-l when i get something working [18:56:20] #action YairRand to propose at wikitech-l next steps for this RfC after Wikimania [18:56:21] brion: great :) [18:56:43] Alright, anything else? [18:57:03] not that I can remember at the moment, hopefully I'm not forgetting something major... [18:57:13] great sounds like we’re about wrapped up [18:57:21] Thank you YairRand for a well prepared RfC meeting [18:57:27] #action brion to report on cssjanus progress to wikitech-l somewhere in the next few weeks [18:57:34] Thank you everybody for your participation [18:57:40] \o/ [18:57:44] Thank you very much :) [18:57:49] #ENDMEETING [18:57:50] Meeting ended Wed Jul 16 18:57:49 2014 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) [18:57:50] Minutes: https://tools.wmflabs.org/meetbot/wikimedia-office/2014/wikimedia-office.2014-07-16-18.02.html [18:57:50] Minutes (text): https://tools.wmflabs.org/meetbot/wikimedia-office/2014/wikimedia-office.2014-07-16-18.02.txt [18:57:50] Minutes (wiki): https://tools.wmflabs.org/meetbot/wikimedia-office/2014/wikimedia-office.2014-07-16-18.02.wiki [18:57:51] Log: https://tools.wmflabs.org/meetbot/wikimedia-office/2014/wikimedia-office.2014-07-16-18.02.log.html [18:57:51] :D [18:58:56] * YairRand leaves with giant smile [18:59:02] :) [18:59:07] yayyyy [18:59:13] good meeting everyone :) see ya next time [19:00:27] qgil: thanks