28265 2015-03-22 00:17:26 +0000 [webvtt] incomplete ruby implementation [I18N-ISSUE-431] 2016-11-17 18:36:57 +0000 1 1 1 Unclassified TextTracks CG WebVTT unspecified PC All RESOLVED MOVED widereview, v2 P2 normal --- 23874 1 silviapfeiffer1 dave.null addison dakim fantasai.bugs fs ishida philipj silviapfeiffer1 singer stillbookish www-international zcorpan public-texttracks oldest_to_newest 118777 0 silviapfeiffer1 2015-03-22 00:17:26 +0000 Feedback by Addison Phillips from W3C I18N group: http://lists.w3.org/Archives/Public/public-tt/2015Mar/0064.html I18N comment: https://www.w3.org/International/track/issues/431 http://www.w3.org/TR/webvtt1/#dfn-webvtt-ruby-object The spec defines a 'ruby' span and an 'rt' span. This is only part of the general ruby model in HTML5. Did you intend to be so specific or were you just looking at old version of html? 118846 1 philipj 2015-03-23 03:54:09 +0000 What is missing? The only other element I can see in HTML is <rp>, which is for "user agents that don't support ruby annotations." It looks like <rp> was added to HTML in 2008: https://html5.org/r/1704 The beginning of WebVTT was in 2010: https://html5.org/r/5079 I think the idea is simply to not cater to UAs that don't support ruby; if you support WebVTT you must support ruby. This seems reasonable to me at least. 118856 2 zcorpan 2015-03-23 08:17:13 +0000 W3C HTML5 also has rb, rtc That said, I think it is a non-goal to match the expressiveness of HTML in WebVTT. WebVTT only has elements to cater for specific use cases that apply to it. 118859 3 silviapfeiffer1 2015-03-23 08:40:34 +0000 See bug 23874 for a discussion on a broader support of ruby. 118864 4 philipj 2015-03-23 10:08:25 +0000 Hmm, OK. Some real-world examples of Japanese subtitles that need something beyond what we currently we have would be useful, if they exist. 120803 5 silviapfeiffer1 2015-06-08 11:20:51 +0000 It seems Firefox have now implemented support for <ruby>, <rb>, <rt>, <rtc> and <rp>, see https://hacks.mozilla.org/2015/03/ruby-support-in-firefox-developer-edition-38/ , so this seems to fully follow http://darobin.github.io/html-ruby/ . We can probably do without <rp>, but should probably include the others. Also see http://www.w3.org/International/tests/repository/html5/the-ruby-element/results-ruby . Do we want to follow the resolution of bug 23874 to move to v2 or include now? 120808 6 ishida 2015-06-08 12:06:50 +0000 (In reply to Silvia Pfeiffer from comment #5) > Also see > http://www.w3.org/International/tests/repository/html5/the-ruby-element/ > results-ruby . Note that, a few days ago, i updated that page in a new location (i'll put a redirect in soon), which shows the latest support in Firefox (and also some improvements in Blink). See http://www.w3.org/International/tests/repo/results/ruby-html 121549 7 dakim 2015-07-01 18:22:57 +0000 There are two very specific use cases that limit the conversion of Japanese subtitles to VTT: https://www.w3.org/Bugs/Public/show_bug.cgi?id=28183 https://www.w3.org/Bugs/Public/show_bug.cgi?id=28184 122530 8 silviapfeiffer1 2015-08-09 10:23:50 +0000 (In reply to Dae Kim from comment #7) > There are two very specific use cases that limit the conversion of Japanese > subtitles to VTT: > > https://www.w3.org/Bugs/Public/show_bug.cgi?id=28183 > https://www.w3.org/Bugs/Public/show_bug.cgi?id=28184 Those two seem a bit premature, but would be easy to add to the allowed CSS markup. What do we do about elements? Add <rb> and <rtc> ? 122552 9 dakim 2015-08-10 22:57:33 +0000 (In reply to Silvia Pfeiffer from comment #8) > (In reply to Dae Kim from comment #7) > > There are two very specific use cases that limit the conversion of Japanese > > subtitles to VTT: > > > > https://www.w3.org/Bugs/Public/show_bug.cgi?id=28183 > > https://www.w3.org/Bugs/Public/show_bug.cgi?id=28184 > > Those two seem a bit premature, but would be easy to add to the allowed CSS > markup. > > What do we do about elements? Add <rb> and <rtc> ? I assumed the <ruby> element was (In reply to Silvia Pfeiffer from comment #8) > (In reply to Dae Kim from comment #7) > > There are two very specific use cases that limit the conversion of Japanese > > subtitles to VTT: > > > > https://www.w3.org/Bugs/Public/show_bug.cgi?id=28183 > > https://www.w3.org/Bugs/Public/show_bug.cgi?id=28184 > > Those two seem a bit premature, but would be easy to add to the allowed CSS > markup. > > What do we do about elements? Add <rb> and <rtc> ? I assumed <ruby> implied <rb> functionality: 00:00:10.000 --> 00:00:15.000 <ruby>Line 1<rt.over>ruby</rt></ruby> <ruby>Line 2<rt.under>ruby2</rt></ruby> But makes sense to add it and would be simple to account for it: 00:00:10.000 --> 00:00:15.000 <ruby><rb>Line 1</rb><rt.over>ruby</rt></ruby> <ruby><rb>Line 2</rb><rt.under>ruby2</rt></ruby> I'm not sure <rbc> or <rtc> is needed at the moment. In terms of practical use, Japanese formats do not use this sort of complex style. Simple ruby support via <rb> and <rt> is sufficient to translate into CSS/VTT. 122600 10 singer 2015-08-13 22:13:01 +0000 I think support for rbc and rtc are not important. My understanding is that rbc and rtc are required either 1) If you want to specify a base character for each segment of ruby, or 2) if you want to have two rubys, one on the right and the other on the left. I believe the use of these complex rubys are rare (if they exist at all) in captions. 122601 11 silviapfeiffer1 2015-08-14 00:04:20 +0000 OK, so all we need to add at this point is <rb> (note: there's no automation in <ruby> implying support of any other element). 122602 12 stillbookish 2015-08-14 00:16:52 +0000 I think that full Ruby support is very important for WebVTT. I think that not having full support is impeding adoption of WebVTT for Japanese subtitles. I think we should adopt HTML Ruby support wholesale into WebVTT. That approach seems simplest. 122604 13 silviapfeiffer1 2015-08-14 05:27:50 +0000 Courtney, do you have any data that supports your statement? What features do subtitles in Japan need? Do you have any examples that would need <rbc> and <rtc> ? 122620 14 stillbookish 2015-08-14 19:13:34 +0000 Silvia, In terms of data, I have requests for support for additional Ruby features from multiple content creators. As I understand it, Japanese subtitles have been delivered most commonly as images , which allows for a great degree of styling and formatting control. And, the expectation is that the same set of functionality will be available with a text-based subtitles delivery format like WebVTT. And until that is the case, there is reluctance to move to WebVTT. Here are some of the features I am being asked for: 1. Ruby alignment, specifically the ability to specify spacing of multi-character Ruby over main text. There are two cases- Ruby that annotates a span of multiple characters in the main text, and Ruby that contains multiple characters. Sometimes Ruby can be applied to individual characters in the main text and sometimes it needs to be applied to a span of them. Support for <rtc> is required to apply Ruby to a span of characters in the main text, and to group Ruby characters together. I think that <rbc> doesn't need to be explicit in most cases. While XHMTL defines <rbc>, it is generated as an anonymous box in HTML; WebVTT could do the same. ( Disclaimer: I'm not an HTML expert, but my reference for that information is the ruby-base-container definition in https://drafts.csswg.org/css-ruby/, and the lack of <rbc> in the HTML spec. I actually can't find where <rbc> is defined, so if someone can send a reference to it, it would be appreciated. ) 2. Ruby placement - under/over ( for horizontal ) and left/right ( for vertical ) of main text. In my understanding, placing Ruby under main text is rare ( though there are examples of it in the HTML spec ). But, with vertical subtitles, being able to specify left or right of the text is valuable. Looks like this can be accomplished with the CSS ruby-position property (https://drafts.csswg.org/css-ruby/) though the spec seems to be a draft. 3. Bōten emphasis dots - these can be delivered as Ruby annotations if the character for them is available. I have not had requests for support for <rp> or for double Ruby annotations-i.e: above and below main text. But, my feeling is that we should probably just adopt the HTML Ruby support wholesale in WebVTT, as the CSS Ruby support proposal is designed to work with it. 122621 15 fantasai.bugs 2015-08-14 20:36:36 +0000 (In reply to Courtney from comment #14) > Silvia, > > Here are some of the features I am being asked for: > 1. Ruby alignment, specifically the ability to specify spacing of > multi-character Ruby over main text. There are two cases- Ruby that > annotates a span of multiple characters in the main text, and Ruby that > contains multiple characters. Sometimes Ruby can be applied to individual > characters in the main text and sometimes it needs to be applied to a span > of them. Support for <rtc> is required to apply Ruby to a span of > characters in the main text, and to group Ruby characters together. I think > that <rbc> doesn't need to be explicit in most cases. While XHMTL defines > <rbc>, it is generated as an anonymous box in HTML; WebVTT could do the > same. ( Disclaimer: I'm not an HTML expert, but my reference for that > information is the ruby-base-container definition in > https://drafts.csswg.org/css-ruby/, and the lack of <rbc> in the HTML spec. > I actually can't find where <rbc> is defined, so if someone can send a > reference to it, it would be appreciated. ) Spanning multiple characters is accomplished by placing more than one character in the <rb>, not by using <rbc> or <rtc>. You are confusing what a "base" is with what a "character" is: in CJK languages each base is frequently one character, but this is not necessary. <rtc> support would be needed only if you have multiple levels of ruby. > 2. Ruby placement - under/over ( for horizontal ) and left/right ( for > vertical ) of main text. In my understanding, placing Ruby under main text > is rare ( though there are examples of it in the HTML spec ). But, with > vertical subtitles, being able to specify left or right of the text is > valuable. Looks like this can be accomplished with the CSS ruby-position > property (https://drafts.csswg.org/css-ruby/) though the spec seems to be a > draft. Yes, ruby-position will do this. There is no HTML support for this since it is a presentational consideration, not a semantic one. > 3. Bōten emphasis dots - these can be delivered as Ruby annotations if the > character for them is available. > > I have not had requests for support for <rp> or for double Ruby > annotations-i.e: above and below main text. But, my feeling is that we > should probably just adopt the HTML Ruby support wholesale in WebVTT, as the > CSS Ruby support proposal is designed to work with it. This should be accomplished by supporting the text-emphasis property: http://www.w3.org/TR/css-text-decor-3/#emphasis-marks 122622 16 dakim 2015-08-15 00:02:25 +0000 Based on Courtney's comment re: "Japanese subtitles allows for a great degree of styling and formatting control. And, the expectation is that the same set of functionality will be available with a text-based subtitles delivery format like WebVTT." We can infer that this issue should be changed from "incomplete ruby implementation" to "incomplete Japanese features support". In which case, we also need to support: https://drafts.csswg.org/css-writing-modes/#text-combine-upright via: https://www.w3.org/Bugs/Public/show_bug.cgi?id=28184 123387 17 stillbookish 2015-09-30 19:26:57 +0000 Here is the link to the portion of HTML5 that defines Ruby tags: http://www.w3.org/TR/html5/text-level-semantics.html#the-ruby-element 123388 18 stillbookish 2015-09-30 19:27:55 +0000 Here is the link to the portion of HTML5 that defines Ruby tags: http://www.w3.org/TR/html5/text-level-semantics.html#the-ruby-element 123389 19 silviapfeiffer1 2015-09-30 19:37:52 +0000 (In reply to fantasai from comment #15) > (In reply to Courtney from comment #14) > > Silvia, > > > > Here are some of the features I am being asked for: > > 1. Ruby alignment, specifically the ability to specify spacing of > > multi-character Ruby over main text. There are two cases- Ruby that > > annotates a span of multiple characters in the main text, and Ruby that > > contains multiple characters. Sometimes Ruby can be applied to individual > > characters in the main text and sometimes it needs to be applied to a span > > of them. Support for <rtc> is required to apply Ruby to a span of > > characters in the main text, and to group Ruby characters together. I think > > that <rbc> doesn't need to be explicit in most cases. While XHMTL defines > > <rbc>, it is generated as an anonymous box in HTML; WebVTT could do the > > same. ( Disclaimer: I'm not an HTML expert, but my reference for that > > information is the ruby-base-container definition in > > https://drafts.csswg.org/css-ruby/, and the lack of <rbc> in the HTML spec. > > I actually can't find where <rbc> is defined, so if someone can send a > > reference to it, it would be appreciated. ) > > Spanning multiple characters is accomplished by placing more than one > character in the <rb>, not by using <rbc> or <rtc>. You are confusing what a > "base" is with what a "character" is: in CJK languages each base is > frequently one character, but this is not necessary. > > <rtc> support would be needed only if you have multiple levels of ruby. Based on this, I am not convinced we need to add any more elements for the WebVTT specific use case. Particularly since Mozilla reported in bug 23874 : "We looked and didn't find any wild examples of <rp> in subtitles." > > 2. Ruby placement - under/over ( for horizontal ) and left/right ( for > > vertical ) of main text. In my understanding, placing Ruby under main text > > is rare ( though there are examples of it in the HTML spec ). But, with > > vertical subtitles, being able to specify left or right of the text is > > valuable. Looks like this can be accomplished with the CSS ruby-position > > property (https://drafts.csswg.org/css-ruby/) though the spec seems to be a > > draft. > > Yes, ruby-position will do this. There is no HTML support for this since it > is a presentational consideration, not a semantic one. It seems the main requirement for ruby extensions on top of what is currently in WebVTT is this ruby-position CSS feature. > > 3. Bōten emphasis dots - these can be delivered as Ruby annotations if the > > character for them is available. > > > > I have not had requests for support for <rp> or for double Ruby > > annotations-i.e: above and below main text. But, my feeling is that we > > should probably just adopt the HTML Ruby support wholesale in WebVTT, as the > > CSS Ruby support proposal is designed to work with it. > > This should be accomplished by supporting the text-emphasis property: > http://www.w3.org/TR/css-text-decor-3/#emphasis-marks OK, we can leave emphasis-marks for the next version of WebVTT. 123391 20 silviapfeiffer1 2015-09-30 19:40:16 +0000 *** Bug 23874 has been marked as a duplicate of this bug. *** 123401 21 silviapfeiffer1 2015-09-30 22:27:33 +0000 After discussion with some internationalisation users from diverse video players at FOMS, we have come to the conclusion that this, together with bug 23874 (rb support), bug 28183 (ruby-position CSS support), bug 28184 (text-combine-upright), and support for CSS ruby-align should be done for the next version of WebVTT, so we can get a interoperable v1 version of WebVTT finalised asap. 123999 22 zcorpan 2015-10-30 04:12:46 +0000 Discussion with ishida at TPAC we concluded that the current subset is OK in webvtt for now, but we should reconsider adopting all of ruby if it gets implemented by all browsers for HTML (including styling, not just HTML parsing). 124220 23 ishida 2015-11-16 12:32:31 +0000 continued at https://github.com/w3c/webvtt/issues/264 127722 24 singer 2016-10-11 17:29:00 +0000 We should re-assess the status of Ruby implementations in all browsers and consider whether we need to adapt VTT in the light of current practice. Re-opening. 127751 25 ishida 2016-10-12 12:44:36 +0000 David, please continue the discussion in the (open) github issue, not by reopening this. 128196 26 singer 2016-11-16 23:10:08 +0000 comment copied from GitHub: After asking experts, I believe we have the following state: <ruby> and <rt>: are in the specification now <rb>: is in by implication (it's a normal part of <ruby> if I understand it correctly), but we should be explicit <rtc>: might be needed for some cases; should we include it? I am not aware of need, myself. <rp>: is only in HTML for backwards compatibility and is not needed. Can we mention <rt> as being part of <ruby> and leave <rtc> for implementer discretion and the future? 128206 27 singer 2016-11-17 18:36:57 +0000 OK, I meant this: mention rB explicitly as being included leave rtc and rbc to future and implementation choice leave out rp (sorry about the typos)