00:00:01 pal: I've not found a technical defect with this condition structure, but I think it's incredibly complex. 00:00:17 ... It forces people to build a parser for the condition language. It's additive and people are already having 00:00:36 ... trouble just rendering TTML. I think it adds a lot of complexity for what originally were a small set of use 00:00:37 ... cases. 00:01:07 glenn: There's a trade-off between expressiveness and complexity. If you look at e.g. media query, which people 00:01:28 ... might quite likely want to conditionalise content based on, it often generates situations where you need 00:01:44 ... at least logical expressions AND, OR, NOT etc and pretty soon want to compare parameters to others, like 00:02:01 ... is width or height < or > some value, then you've got numerical comparisons as well as logical ones. Then 00:02:23 ... soon you want arithmetic too! Just adding conditions that have named parameters and combinatorial logic 00:02:37 ... for logical expressions doesn't seem very different to me than adding something that satisfies the more 00:02:53 ... general case, and from a parsing perspective it's really straightforward. To write a parser for what is there 00:03:17 ... now vs a subset is going to have almost the same complexity. The syntax now will require a lot of test content though. 00:04:06 ... I didn't want to reinvent media queries. To implement the expression language I'd estimate as a day's work. 00:06:00 nigel: You may want to conditionalise XML attributes on the tt:tt element but I can't see how you'd do that. 00:06:17 ... For example you might want the extent to be dependent on whether the media is being viewed fullscreen 00:06:32 ... or not fullscreen. 00:07:13 glenn: There are limits to it. The precedent for this is the switch element in SMIL and SVG. 00:07:25 ... What that system did was to limit the locations where switch is permitted, e.g. you can't use it to switch 00:07:48 ... between definitions based on conditions. In TTML we have a number of definitional mechanisms like the 00:08:08 ... region and style elements. So far I've permitted condition to be used on those, but the only one it can't 00:08:31 ... apply to is the tt element itself. Is there any way to express the information on tt in a conditional way? 00:08:42 ... I haven't thought about that. 00:09:02 ... I did put condition on tt:tt but that may be overly aggressive. 00:09:11 nigel: You'd have to duplicate the whole tree. 00:09:41 glenn: I'm not sure condition is going to work at all on the tt element. 00:18:54 nigel: It's also unclear to me when the evaluation time is for conditions - is it any time during the presentation? 00:19:30 glenn: My intention was to evaluate once before presentation and not again, but I guess implementations could 00:19:49 ... offer on-the-go reevaluation. That could be implementation dependent. 00:20:12 nigel: In terms of the spec it's not clear when the evaluation time should be. 00:20:24 glenn: Certainly adding a note that needs to be defined would be prudent. 00:21:54 nigel: You could also take an approach of using XML Query or other syntax and re-use existing techniques 00:22:04 ... as a single evaluation before presentation. 00:22:35 glenn: If you're suggesting throwing this out and using XML Query instead I think that would be overkill. 00:23:01 ... I think what's in scope of our discussion is 'is there a testable implementable use case' that this meets, rather 00:23:24 ... than complexity of implementation. Complexity is a profile issue. Let's say you want a profile that uses 00:23:49 ... condition and limits the form of the expressions, such as 'the only permitted functions are bound parameters' 00:24:11 tai: I have recently been wondering if there will ever be a complete TTML implementation. 00:25:00 ... You always need something like IMSC to be an implementable thing. 00:25:30 tai: One proposal for this conditional mechanism: a complete example and use case in the spec would make it 00:25:33 ... easier to understand. 00:25:45 glenn: I agree it needs some examples. 00:27:28 nigel: What's the list of use cases we have for this so far: forced display, ...? 00:27:33 glenn: languages 00:28:20 Courtney: I get asked for functionality to present translated subtitles as a licence requirement in some countries, tied to the audio language. 00:28:58 ... e.g. if audio language == spanish then you must display subtitles in language XYZ. This is very similar 00:29:02 ... to forced subtitles. 00:29:34 tai: I thought of a use case that may or may not work, where you have a complete style set that is in use or 00:31:54 ... out of use depending on the condition, to select predefined style sets. 00:32:00 nigel: That could help meet the MAUR requirements 00:32:08 pal: It wouldn't work for regulation. 00:32:29 Courtney: full customisation at the terminal is a requirement for FCC 00:36:36 pal: We don't need to put that functionality into TTML - it's a receiver thing only. 00:37:55 nigel: Maybe it wouldn't work in the US but based on the feedback I see to BBC I think it would really be 00:38:12 ... appreciated if broadcasters could offer different style sets, and not outrageous for broadcasters to provide 00:38:15 ... them. 00:38:39 tai: In Germany I can see that a useful option might be to put this in the document. We have discussions 00:39:02 ... with TV manufacturers too and they don't see it as their business to provide configuration options. 00:40:03 ... It's unclear how customisation works now - it has to be done by content provider apps. 00:46:03 Courtney: Whatever layer is doing the rendering is where the options have to be evaluated. 00:46:34 nigel: You could have some content provider provided transformations to allow for some styling. 00:46:49 ... There's a problem with any kind of declarative styling which is that to style specific content you need to 00:47:13 ... know how it's identified so that you can target (or select) it. Only the document author knows that, because 00:47:36 ... it's document dependent. 00:48:14 ... That makes customisation always difficult. 00:48:36 Courtney: It could be interesting to investigate whether pre-defined transformations can meet all the user 00:48:50 ... requirements for accessibility but there would be lots of work to do on it. 00:49:33 glenn: We have to leave complexity until we've seen the implementations, unless someone can say what 00:49:37 ... doesn't work about it. 00:49:57 ... Obviously conditionalisation on the tt element is a limiting factor. Can we mitigate that and should we mitigate that? 00:50:25 ... If the solution is very complex to specify and hard to understand then that could count against it. 00:50:54 ... My main criteria are: is it specifiable, is it usable, is it implementable? Equally important. 00:51:16 ... By the way we need to take out stereoLeft and stereoLeft. 00:51:31 Action: glenn remove stereoLeft and stereoRight from 00:51:31 Created ACTION-385 - Remove stereoleft and stereoright from [on Glenn Adams - due 2015-04-17]. 00:52:50 glenn: Are there any other bound parameters that folk want? 00:53:02 nigel: You could add full-screen and then we can see how that plays out. 00:57:47 nigel: The difference would be that for full-screen you want extent="contain" but if not then "cover" because 00:57:54 ... there's more display to use. 00:58:23 glenn: Let me think about that bigger level question there. I've already got a thought process in train for 00:58:47 ... how to handle conditionalised parameters. In TTML2 we also allow all the style attributes to be added onto 00:59:20 ... the tt:tt element and have them inherited by the region elements. They would maybe also need to be 00:59:24 ... conditonal. 00:59:37 nigel: But what's the difference between that and the initial element? 00:59:52 glenn: I guess if you specified initial differently and did not specify them on a region then region would pick 01:00:04 ... them up. Yes, maybe its redundant to put them on tt. 01:00:32 ACTION: glenn to investigate whether region style inheritance from tt is necessary given the ability to redefine initial values 01:00:32 Created ACTION-386 - Investigate whether region style inheritance from tt is necessary given the ability to redefine initial values [on Glenn Adams - due 2015-04-17]. 01:02:20 Topic: Recap 01:02:56 nigel: We've covered a lot today. Tomorrow we'll use the 'spare' time available to go over the remaining 01:03:25 ... TTML2 topics as well as everything else we have planned - let's prioritise based on time available and what 01:03:32 ... everyone wants to do. 01:03:55 ... We're restarting in the morning at 8:30. 01:03:58 rrsagent, generate minutes 01:03:58 I have made the request to generate http://www.w3.org/2015/04/10-tt-minutes.html nigel 01:04:19 nigel: Thanks everynone! [adjourns meeting] 01:04:37 rrsagent, generate minutes 01:04:37 I have made the request to generate http://www.w3.org/2015/04/10-tt-minutes.html nigel 01:04:47 glenn has joined #tt 01:07:09 rrsagent, generate minutes 01:07:09 I have made the request to generate http://www.w3.org/2015/04/10-tt-minutes.html nigel 01:10:02 rrsagent, this meeting spans midnight 01:10:06 rrsagent, generate minutes 01:10:06 I have made the request to generate http://www.w3.org/2015/04/10-tt-minutes.html nigel 01:37:45 nigel has joined #tt 03:52:07 glenn has joined #tt 15:25:25 RRSAgent has joined #tt 15:25:25 logging to http://www.w3.org/2015/04/10-tt-irc 15:25:27 RRSAgent, make logs public 15:25:27 Zakim has joined #tt 15:25:29 Zakim, this will be TTML 15:25:29 I do not see a conference matching that name scheduled within the next hour, trackbot 15:25:30 Meeting: Timed Text Working Group Teleconference 15:25:30 Date: 10 April 2015 15:25:46 chair: nigel 15:26:01 Present: glenn, pal, tmichel, nigel, Courtney, tai 15:26:05 scribe: nigel 15:26:14 rrsagent, this meeting spans midnight 15:33:51 Topic: Agenda today 15:35:00 nigel: Yesterday's minutes are visible at http://www.w3.org/2015/04/09-tt-minutes.html 15:36:00 tmichel has joined #tt 15:36:12 dakim has joined #tt 15:39:46 nigel_ has joined #tt 15:40:01 glenn: Yes, I've been thinking about that too. I'll create an issue for it. 15:41:04 nigel: tts:ipd can apply to div, p and span but tts:textAlign only applies to p and span so maybe its needed there. 15:41:47 glenn: I'll create an issue for that too. 15:43:21 issue-384? 15:43:21 issue-384 -- tts:displayAlign may need to apply to content elements -- raised 15:43:21 http://www.w3.org/AudioVideo/TT/tracker/issues/384 15:43:26 reopen issue-384 15:43:26 Re-opened issue-384. 15:44:23 issue-385? 15:44:23 issue-385 -- tts:textAlign may need to apply to div, body -- raised 15:44:23 http://www.w3.org/AudioVideo/TT/tracker/issues/385 15:44:29 reopen issue-385 15:44:29 Re-opened issue-385. 15:45:23 nigel: My last thought on this one is that the empty spans with ipd and bpd shouldn't be pruned. 15:45:30 ... This is related to issue-368 15:45:33 issue-368? 15:45:33 issue-368 -- The [Construct Intermediate Document] process erroneously prunes empty
elements -- open 15:45:33 http://www.w3.org/AudioVideo/TT/tracker/issues/368 15:46:23 issue-368: [Meeting 2015-04-10] Any empty content element that generates an area e.g. by setting ipd or bpd must not be pruned either. 15:46:23 Notes added to issue-368 The [Construct Intermediate Document] process erroneously prunes empty
elements. 15:46:46 nigel: As there are no more thoughts from yesterday, let's crack on with the planned agenda. 15:47:12 Topic: IMSC 1 Test Suite and Implementations 15:48:04 https://www.w3.org/wiki/TimedText/IMSC1_Implementation_Report#Tests 15:48:21 pal: The test suite seems pretty stable. I know we've had some comments back that have been corrected. 15:48:43 ... As far as I know its available for people to test. I was recently at an event where a couple of people 15:49:11 ... tried to run the test suite for documents wrapped into MXF. I think people are trying it and experimenting. 15:49:11 courtney has joined #tt 15:49:17 tai: Do they run the TTML tests too? 15:49:54 pal: I don't know about that - but the maximum rate tests do exercise a lot of feature areas. 15:50:03 ... I'm focusing on the IMSC 1 tests. 15:50:40 ... I know that some folk have tried this but getting them to contribute their tests to W3C is a challenge. 15:50:58 ... They're not quite sure how the work is going to be used. 15:51:28 nigel: So maybe we need some better comms here to help people understand. 15:51:43 pal: A lot of implementors will not be W3C members because the web isn't their main business. 15:51:48 ... So how do we manage this? 15:52:52 nigel: Another data point here is an example I had recently - I was talking to someone whose dev team 15:53:05 ... was in another part of the world and they didn't feel they could push that team to do extra work, but 15:53:19 ... when they had delivered they would be able to try it. So maybe its lead time too. 15:54:09 glenn: For TTML1 we had a template spreadsheet. 15:54:51 tai: For manufacturers to proceed without guidance can be quite difficult. There may be a win win if 15:55:09 ... we can also help them with their implementation work, in providing guidance, advice etc. on the spec. 15:55:29 pal: Another obstacle that's related but different is that some manufacturers have as a policy non-disclosure 15:56:03 ... of results until a spec is published, to avoid appearing to have non-compatibility with the spec until it has 15:56:09 ... been formally finished. 15:56:56 tmichel: You don't necessarily have to make public the implementor's name - they can be anonymised, as long 15:57:02 ... as we know in the group which one is which. 15:57:19 glenn: I heard you use the word "compatibility" which they don't have to claim, for our purposes. 15:57:27 pal: That's a good point. 15:57:45 ... W3C has implementation before standardisation - that's different from other groups where they 15:57:53 ... standardise, implement and then fix the standard if needed. 15:58:17 tai: We do see that in e.g. XML too where products area available before the spec is a recommendation. 15:58:57 ... For commercial vendors if you give clear guidance and offer work together that would help. 15:59:06 ... Another possibility is to see if there's an open source implementation. 15:59:12 pal: That would be ideal. 15:59:36 tai: In my experience with open source groups they are quite keen and do want real example content as 15:59:39 ... well as test content. 16:00:26 ... For presentation, that would be more a player like VLC for example. IRT has been in touch with them as 16:00:44 ... part of the HBB4ALL group, and when we discussed with them the EBU-TT-D standard and the other related 16:01:00 ... standards the VLC groups were quite interested in implementing when there is test content available. 16:01:15 ... They gave me the impression they might implement based on examples rather than the spec. 16:01:24 dakim: Caption vendors always ask for an xsd. 16:01:45 pal: If you give them the maximum rate subtitle stuff, that has some useful coverage of styles, timing etc. 16:02:25 nigel: Have we issued a formal invitation to provide implementations? 16:02:41 tmichel: We did that formally when we published CR. 16:02:45 nigel: Who saw that? 16:02:57 tmichel: W3 groups, members. It was on the W3 home page. 16:03:08 pal: A lot of the implementors won't have seen that. 16:03:24 tmichel: If we know who we want to target I can help send the comms to them. 16:03:31 glenn: It's best not to ask for general input. 16:03:55 pal: I really like the idea of a form. 16:03:58 dakim_ has joined #tt 16:04:07 nigel: Like a WBS form that's open to non-members? 16:04:17 pal: Could be, or even a PDF form. 16:05:03 nigel: I think there are two actions here: 1. Create the set of questions; 2. Collate the list of recipients. 16:05:22 pal: I'd be happy to provide the list as I know it. I'd like some help with the form, since something similar 16:05:25 ... has already been done. 16:12:58 tai: What does this mean for the schedule? 16:13:19 tmichel: I was proposing to start by contacting potential implementors by email, and then maybe invite 16:13:40 ... them to attend a telecon. Another thing is to invite people as invited expert during the implementation 16:13:42 ... period. 16:14:00 nigel: I think that sounds a bit heavyweight - we don't need the invited expert bit IMO. 16:14:19 tai: It's important not to make joining as an invited expert a requirement because the process isn't so easy, to join. 16:14:53 pal: Does W3C have a presence at IBC? 16:14:57 tai: EBU does. 16:15:03 glenn: W3C has a presence at NAB. 16:15:10 pal: That's a little soon, being tomorrow! 16:15:32 tai: There may well be presentations at the IBC EBU booth - if you're there we can check with the EBU group and 16:15:52 ... Frans if we can mention IMSC too. We get feedback that there's confusion about how all the different standards 16:16:03 ... relate to each other. 16:16:11 nigel: Sounds good to me. 16:16:47 s/me./me (wearing my EBU group co-chair hat). 16:17:00 Action: pal Collate list of potential implementors of IMSC 1. 16:17:00 Created ACTION-387 - Collate list of potential implementors of imsc 1. [on Pierre-Anthony Lemieux - due 2015-04-17]. 16:17:24 Action: tmichel Start preparing the IMSC 1 Implementation report form. 16:17:25 Created ACTION-388 - Start preparing the imsc 1 implementation report form. [on Thierry Michel - due 2015-04-17]. 16:18:08 Action: tai Check with EBU group if we can have a combined event at IBC with EBU group and IMSC/W3C 16:18:09 Created ACTION-389 - Check with ebu group if we can have a combined event at ibc with ebu group and imsc/w3c [on Andreas Tai - due 2015-04-17]. 16:22:54 dakim has joined #tt 16:24:13 nigel: I found a couple of issues from a colleague who implemented the linePadding and multiRowAlign tests 16:24:17 ... and from Andreas. 16:24:44 ... 1. The example PNGs don't honour the displayAlig="center" setting (because I forgot that in the HTML) 16:24:57 ... 2. The region doesn't begin 10% down 16:26:16 ... 3. The text is so long that in real world implementations it doesn't necessarily fit on a line. 16:26:50 nigel: I think we need to regenerate the PNGs to match the TTML, and also make the text shorter. 16:27:25 Action: nigel Update the multiRowAlign and linePadding TTML files to have shorter lines of text 16:27:25 Created ACTION-390 - Update the multirowalign and linepadding ttml files to have shorter lines of text [on Nigel Megitt - due 2015-04-17]. 16:27:27 Action: nigel Update the multiRowAlign and linePadding PNGs to match the TTML 16:27:28 Created ACTION-391 - Update the multirowalign and linepadding pngs to match the ttml [on Nigel Megitt - due 2015-04-17]. 16:28:24 tai: I also noticed that the document order of the

elements is different from the display order of the regions 16:28:43 ... So the first subtitle with the green background is rendered at the bottom and the second subtitle with 16:28:59 ... the black background is rendered at the top. I think it would confuse people so I'd suggest to change that. 16:29:11 ... It may just be a distraction. 16:30:37 pal: I'll be happy to change that. 16:30:52 nigel: This came from pal's original examples. I didn't change it because although it looks surprising it isn't 16:30:58 ... wrong. I almost quite like it! 16:32:05 Action-390: Also change the region ordering to be more intuitive. 16:32:05 Notes added to Action-390 Update the multirowalign and linepadding ttml files to have shorter lines of text. 16:32:22 nigel: Some of the other example have it too, e.g. forcedDisplay1.ttml 16:32:41 Action: pal tweak the examples with multiple regions so that they appear in a more intuitive order. 16:32:41 Created ACTION-392 - Tweak the examples with multiple regions so that they appear in a more intuitive order. [on Pierre-Anthony Lemieux - due 2015-04-17]. 16:33:02 Action-392: (nigel will do the multiRowAlign and linePadding ones as per Action-390) 16:33:02 Notes added to Action-392 Tweak the examples with multiple regions so that they appear in a more intuitive order.. 16:35:40 nigel: I'd like to show a prototype implementation that my colleagues at BBC R&D have put together based on 16:36:16 ... gstreamer, using MPEG DASH to publish the content. I asked them to put the multiRowAlign and linePadding 16:36:35 ... examples through it. They had to make a tweak because they're not actually valid EBU-TT-D as they use 16:37:02 ... time expressions like "1s" where EBU-TT-D requires "hh:mm:ss.sss" type format. 16:37:10 pal: Okay let's change that in the examples. 16:38:24 Action-390: Change the time expressions to EBU-TT-D compliant ones too. 16:38:24 Notes added to Action-390 Update the multirowalign and linepadding ttml files to have shorter lines of text. 16:39:02 nigel: There are some tests that could be EBU-TT-D compatible too with these changes. 16:40:18 courtney_ has joined #tt 16:40:21 Action-392: Also change the time expressions to be EBU-TT-D compliant on all that could be compliant 16:40:21 Notes added to Action-392 Tweak the examples with multiple regions so that they appear in a more intuitive order.. 16:43:20 pal: Can we add this to the implementation report? 16:43:47 nigel: Yes, as far as I'm concerned we can do that. Thierry is there any reason why not? 16:44:08 tmichel: No reason why not. It would be good to post the videos too. 16:44:16 nigel: They're large files, but otherwise no problem. 16:44:33 tai: So for the BBC implementation they could just submit that they've passed those two tests? 16:44:53 tmichel: Yes, that's fine. And it's okay to show even one test being passed by any implementation. Each 16:45:07 ... test does need to have at least 2 implementations. 16:45:26 courtney has joined #tt 16:45:26 tai: Is there any requirement to double check that the claims of passing the test are true? 16:45:42 tmichel: We've relaxed this. PReviously we would show the implementations working, but now it's based on 16:45:58 ... trust. THe only thing we should avoid is a misunderstanding of the output that you're supposed to get. You 16:46:15 ... have to compare the output with some proof - it could be a video, a textual description, an image, or an SVG 16:46:20 ... file etc. 16:46:44 tai: Or it could be a member verification? 16:46:55 pal: Yes, I was just asking nigel if BBC could submit this. 16:47:21 tmichel: Each company should add their entry on the implementation report, using the form we're going to make. 16:47:48 pal: There's space on the wiki so until we have the form we could just add a table to the results section 16:48:11 ... on the implementation report, and then we can update when we have a fancier system. 16:48:36 Action: nigel Send pal details of implementation and tests passed so he can create a table in the implementation report 16:48:36 Created ACTION-393 - Send pal details of implementation and tests passed so he can create a table in the implementation report [on Nigel Megitt - due 2015-04-17]. 16:49:48 nigel: So we have a good news story - at least one implementation of two of the tests. 16:51:02 Topic: IMSC 1 schedule and CR publication 16:51:31 rrsagent, generate minutes 16:51:31 I have made the request to generate http://www.w3.org/2015/04/10-tt-minutes.html nigel 16:57:01 courtney has joined #tt 17:04:08 dakim has joined #tt 17:15:07 courtney has joined #tt 17:17:58 pal: Let's look at today's schedule. 17:18:40 https://www.w3.org/wiki/TimedText/Publications 17:19:02 pal: According to our schedule interop testing would have ended and we'd be publishing a CR2 around now. 17:19:21 ... Interop testing is ongoing, and we have made some changes to the spec. Should we have a second CR? 17:20:13 ... Should we proceed with that now, or wait for interop testing and feedback. Do we need a second CR at all? 17:20:28 ... If we decide on a second CR we should maybe just do it according to our schedule. 17:20:40 tmichel: I think it's better to publish a second CR now because there are no implementations yet and so its 17:20:59 ... easier to justify the changes to the director. If you wait until PR then it's a tougher sell. 17:21:07 pal: So maybe that answers that question! 17:21:26 tmichel: The second issue is, if we publish a CR2, we might start thinking about at risk features if there are 17:21:37 ... any, that we may want to remove to make going to PR easier. 17:21:51 nigel: We currently have no at risk features marked. 17:22:10 pal: That's right. The three that were in my mind most "at risk" due to implementations were multiRowAlign, 17:22:25 ... linePadding and image. We've seen an implementation of multiRowAlign and linePadding this morning so 17:22:50 ... they're not at risk, and I've seen plans for doing image already. So I don't believe there should be any at 17:22:52 ... risk features. 17:23:50 nigel: You've created the CR diff. Are there any open issues? 17:24:24 pal: The only open issue is the one on begin, that we opened yesterday. Issue-382. 17:24:41 pal: Based on our discussion I think the answer is clear but I want to check on that some more. 17:24:57 ... We also have issue-375. 17:24:59 issue-375? 17:24:59 issue-375 -- Use of the ebutts:linePadding and ebutts:multiRowAlign in IMSC 1 -- pending review 17:24:59 http://www.w3.org/AudioVideo/TT/tracker/issues/375 17:26:01 pal: In the CR it used to say "may be used" and now has "with the following provisions" and there's also the 17:26:04 ... note added. 17:26:43 tai: After we filed the first issue we found some additional points. Originally we wanted to allow the extension 17:26:57 ... attributes to be permitted on the content elements, in contrast to EBU-TT-D that only allows them on style 17:27:04 ... elements. 17:27:26 ... There was a minor thing about the wording of the note to make clear that there is a difference between 17:27:43 ... IMSC and EBU-TT-D and I proposed some wording that Pierre has added. 17:28:54 tai: I've read the updated text and am happy with it. There may be a small grammar issue - adding a 'the'. 17:28:58 nigel: I've read it too. 17:29:06 pal: I'm going to make that grammar change right now. 17:29:59 dakim: What's the plan for referencing TTML2 from IMSC? 17:30:17 pal: The plan is to make an IMSC 2 that references TTML2. Since both are similarly delayed it makes sense to 17:30:28 ... stay with that plan. Let's discuss more after we're done with IMSC 1 here. 17:32:02 pal: I've pushed the grammar change now. 17:32:18 close issue-375 17:32:18 Closed issue-375. 17:32:29 nigel: Great, now we have only one open issue. 17:32:49 pal: So it sounds like the overhead for a second CR is low based on tmichel's input. 17:33:01 tmichel: Yes, let's not say that these are substantial changes. 17:33:29 pal: Yes. Some of them are normative provisions. Are they extensive or substantially affect the overall scope... 17:33:47 glenn: Historically in this group our criteria has been: does it change the syntax in any way? If yes, it's substantive. 17:34:02 ... If it significantly alters the compliance rules then we've also considered it substantive. 17:34:05 nigel: What does that mean? 17:34:24 glenn: There's some subjectivity in the word "significantly", which is a group choice. 17:34:31 pal: So let's review the changes and determine that. 17:35:03 glenn: We've also got some changes under the wire by saying that they fix a mistake, or haven't been implemented etc. 17:37:30 http://services.w3.org/htmldiff?doc1=http%3A%2F%2Fwww.w3.org%2FTR%2Fttml-imsc1%2F&doc2=https%3A%2F%2Fdvcs.w3.org%2Fhg%2Fttml%2Fraw-file%2Ftip%2Fttml-ww-profiles%2Fttml-ww-profiles.html#features 17:38:13 pal: §6.6 Removal of frame rate relationship with related video object. One of the issues here is that you can't 17:38:31 ... realistically test this without some video. In other words the impact is nil because we couldn't test this anyway. 17:38:58 ... Any implementations that enforce the old language are still compliant. So the practical impact of this 17:39:04 ... change is practically nil. 17:40:21 atai has joined #tt 17:43:59 tmichel: the exact wording in the W3C process defines substantive in http://www.w3.org/2014/Process-20140801/#substantive-change 17:45:34 glenn: In the past we have not considered the 3rd bullet in point 3 above to be substantive. 17:45:43 tmichel: This is the new process though. 17:46:11 pal: By the letter here any change that affects a normative provision falls into point 3. 17:46:41 tai: What's the consequence if some changes are substantive? 17:46:48 tmichel: In that case we need to have a director call. 17:47:43 nigel: Even if we have some substantive changes I'm happy to take the Director through them. 17:50:35 nigel: This specific change to §6.6 falls under the third bullet in point 3, so it does affect conformance; 17:50:40 Zakim has left #tt 17:50:48 ... therefore under the new process it does count as substantive. 17:51:17 nigel: The next one is in §6.7.1. 17:51:34 pal: Again this is relaxing a constraint that used to bind a value in the TTML with an external object. 17:52:04 pal: The first change in this section is in the same category as the previous one we just discussed. 17:52:12 nigel: Right, so this is classed as substantive. 17:52:31 pal: The next one, removing the conformance regarding aspectRatio and extent is substantive for sure 17:52:41 ... because it removes an enforceable check in the document. 17:53:00 nigel: The third change in this section is the removal of a note. 17:53:05 pal: That one is editorial by definition. 17:53:39 nigel: The next change is in §7.4. This changes the initial value of tts:displayAlign. 17:53:48 pal: That would affect presentation, i.e. processor behaviour. 17:54:04 nigel: This is also classed as substantive. 17:54:16 nigel: Next one is the same for textAlign. 17:54:21 nigel: This is also classed as substantive. 17:54:40 nigel: The next one is the set of changes to linePadding and multiRowAlign. 17:55:12 pal: I think they have to be substantive because they clarify the situation relative to the previous version. 17:55:32 nigel: Agreed. 17:55:53 pal: I'd like the group believe that they are extensive enough to warrant doing wide review again for example. 17:56:09 ... In my opinion the answer is no, but maybe we can have a group consensus to take to the Director. 17:56:19 nigel: This is also classed as substantive. 17:56:50 nigel: That was the last change. 17:57:32 nigel: On the point about wide review, some of the changes are relative to SMPTE-TT so arguably we should 17:57:37 ... give them the chance to respond. 17:58:08 ... Just sending to SMPTE is narrow review though not wide review. 17:58:56 nigel: My view is that this doesn't need wide review. One of the changes was in response to the previous 17:58:58 ... wide review. 17:59:14 pal: Yes, that arrived beyond the feedback deadline and we chose to defer knowing that we might have other 17:59:15 ... changes. 18:00:12 nigel: When we've completed the one remaining open action I'll raise a proposal in a future meeting to request 18:00:20 ... transition to CR for the updated ED. 18:04:55 Topic: Steps for exiting CR for IMSC 1 18:05:16 pal: Thierry, what are the steps for exiting, once we've published CR2? 18:05:24 tmichel: First we have to publish CR2! 18:06:10 tmichel: We have to organise the request and the Director's call because there are substantive changes. 18:06:23 pal: I will come back with a proposal for closing the last remaining issue. 18:07:35 nigel: In the CR we have exit criteria, which are just about implementation. For CR2 we will need a new 'earliest 18:07:38 ... date'? 18:07:46 tmichel: Yes, at least 4 weeks beyond CR2 publication. 18:08:17 nigel: So I think the steps are: 1) Complete the implementation report to meet the criteria, 2) request transition to PR. 18:08:26 pal: And we have actions from earlier to help with 1. 18:09:37 pal: On the timeline, we will hopefully do CR2 in April, and probably not make PR in June. While that's going 18:09:47 ... on we can start drafting IMSC 2. 18:10:25 pal: The only thing that will be tough is that SMPTE wanted IMSC 1 to reach Rec in July so they can reference it. 18:10:29 nigel: That's still possible. 18:10:45 pal: True. 18:11:00 nigel: If SMPTE has that goal and can encourage implementations then that will help them. 18:12:02 pal: The plan for IMSC 2 is to base it on IMSC 1 and add any TTML 2 features such as Ruby or SMPTE requested 18:12:07 ... features that people want. 18:13:01 Present+ loretta 18:13:17 Loretta has joined #tt 18:13:26 nigel: [5 minute break] 18:22:21 courtney has joined #tt 18:30:43 Loretta has joined #tt 18:31:03 Topic: WebVTT FPWD Feedback handling 18:33:05 Loretta: How do the CG and the WG work together? 18:33:20 nigel: For Rec track stuff that has to come from the WG and feedback to go back to the WG. 18:33:51 ... If there are any technical changes to the spec needed then it would be bad if the CG isn't happy with them. 18:34:13 Courtney: It can be tricky to get consensus from the CG. 18:34:43 tmichel: The WG has done a wide review. We received comments from i18n and CSS WG. We're expecting 18:35:07 ... review from accessibility. That will close the debate on the wide review, as it will represent feedback from 18:35:13 ... all the groups we have dependencies with. 18:35:26 ... The only group missing is HTML and I'm expecting a statement that they're satisfied. 18:36:14 nigel: I want to explore that some more - for IMSC 1 we took the view that wide review goes beyond just 18:36:26 ... W3C group dependencies and to the wider world. 18:36:40 tmichel: That's even better if we can demonstrate that review, but it's not essential. 18:39:18 nigel: I was expecting more. For example to include the liaisons with external groups. 18:40:16 tai: David Singer did post on the EBU group asking for wide review. That group hasn't provided any formal feedback. 18:40:36 nigel: As a statement of fact, I don't believe there has been any feedback from any external group. I don't have 18:40:43 ... visibility of any other requests David may have made. 18:41:29 http://www.w3.org/2014/Process-20140801/#doc-reviews 18:44:07 nigel: We had a conversation by email about the process for responding and collecting wide review dispositions. 18:44:08 https://lists.w3.org/Archives/Public/public-tt/2015Apr/0006.html 18:44:21 Loretta: who has the action to take that forward? 18:45:36 tai: I think Dave offered to do it, on April 2. 18:45:50 nigel: That email, https://lists.w3.org/Archives/Public/public-tt/2015Apr/0003.html says he doesn't 18:46:01 ... see a problem with doing it, but that's not the same as taking the action, right? 18:46:39 nigel: He offered to do bridging, and my follow-up was attempting to isolate where the bridging is needed. 18:46:52 Courtney: I think what he means is he's asked Silvia to do it. 18:47:31 ... There's another email thread about CSS WG feedback, where a more concise version of it. 18:48:38 Courtney: We should have an ETA and know who is writing the dispositions document we discussed last week. 18:48:49 Loretta: presumably the CG is blocking that? 18:49:04 tmichel: The bridging also includes acceptance by this group of any proposals made by the CG. 18:49:20 pal: In the case of IMSC 1 step 1 was to document every comment, which tmichel did. Then a proposed 18:49:35 ... response was created and the commenter accepted or not and we discussed it. That's exactly what's 18:49:58 ... required by the Director. The first step would be to document every comment received. 18:50:22 tai: The working model between CG and WG is new as tmichel noted last week. 18:50:32 nigel: I agree on both tai's and pal's points and don't think they're exclusive. 18:50:44 pal: So the issues could be in the bug tracker. 18:50:55 Loretta: Can we tag that they're comments so we can formally collate them? 18:51:19 pal: There's also the recurring issue of IP - if the CG creates new IP in the course of responding to the comments, 18:51:40 ... how do we ensure that the IP meets the requirements of W3C, in a continuous way? 18:52:18 nigel: If there's a spec change made by a CG member who is not a WG member then the IP needs to be cleared 18:52:23 ... on a case by case basis. 18:52:47 Courtney: Is this only what's in the spec? 18:52:48 tmichel: Only the spec changes need to be cleared. 18:53:01 Courtney: If the IP question is based on versions of the spec... 18:53:09 pal: Then you might want to do another clearance call. 18:53:24 tmichel: This is so complicated! I've been tracking people from the CG that have been active contributors and 18:53:37 ... some never responded and are listed in the acknowledgement section as active contributors. 18:54:00 Courtney: Part of the problem is that it's taken so long that people may have moved on. 18:54:39 ... It sounds like we need to talk to Dave and maybe Silvia and get clarity on what everyone needs to be produced 18:54:52 ... and a date for when it will be done and who is doing it. Let's get a schedule and then work to it. 18:55:25 tmichel: Now it's in the tracker the CG should make proposals for the comment, as a first step IMO. The 18:55:39 ... second step is for the TTWG to approve them. 18:55:59 Courtney: We don't need to do that piecemeal for every comment. 18:56:17 nigel: Part of the problem is we didn't set a review feedback end date on the FPWD of WebVTT so when do we stop? 18:57:07 Courtney: We can get a specific date for accessibility review. 18:57:09 tmichel: I'll do that. 18:57:27 Courtney: Then we have to make a decision point - do we have wide review? 18:59:41 nigel: We should check the process too on that. I have reservations on this. Let's review the process and charter. 19:00:14 nigel: The process wants the charter's dependencies specifically to have been contacted. I don't think there's 19:00:51 ... any archived evidence of Dave emailing them all, just the W3C groups in 3.1 of the charter. 3.2 and 3.3 19:01:06 ... seem to be missing. The general public has been told, due to publication on the W3C homepage. 19:01:41 ... In terms of appropriate times, the end date for feedback was not clear on the FPWD. 19:02:02 nigel: One way to proceed here is to issue a second WD and a specific call for wide review with an end date 19:02:18 ... and process any feedback from that, having solicited from all the dependencies on the charter. 19:02:28 tmichel: That would be a good way to proceed. 19:02:52 Courtney: I think that sounds like a good set of next steps for getting the wide review done. What's next after 19:02:54 ... that? 19:03:11 tmichel: Then we move to CR. 19:03:35 nigel: Part of the CR is to specify exit criteria, which must include implementation experience. 19:03:54 tmichel: In the pipe we also have issues regarding styling for example, such as inline or external stylesheets. 19:04:07 ... Do we want to have those resolved in the current version or a later version? 19:04:18 Courtney: IMO we should defer to a later version. 19:04:23 Loretta: I would really like it in this version. 19:05:32 Courtney: Then we have to get implementations. I see the value of having it in there but we don't have 19:05:37 ... consensus on the design. 19:05:55 Loretta: To my mind the conversation on this happens in the CG, and I think I saw a consensus emerging. 19:06:14 Courtney: It's been quiet in the CG for the last 6 weeks. 19:06:27 tmichel: Even if we went to CR for the current version I don't think there's evidence of implementation for 19:06:34 ... the whole spec. 19:06:53 Courtney: I don't agree. We've been supporting this in Apple since last fall in two separate implementations. 19:07:04 ... In addition Google has an implementation. 19:07:40 Loretta: It's not complete yet. The parsers and encoders support it, but Chrome does not. It doesn't do regions. 19:07:48 ... What counts as an implementation? 19:08:01 nigel: We can choose that in the CR exit criteria. We have some flexibility. 19:08:21 tmichel: We also need test suites. In my experience browser implementations don't implement a lot of features 19:08:23 ... completely. 19:08:45 tai: This is the same conversation as we had for IMSC, and we decided that we don't need single implementations 19:09:01 ... that pass all the tests, just that every test is passed by at least two implementations. 19:09:14 ... Silvia has started on the test suite in git. 19:09:36 Courtney: We've been working on the tests too but don't have anything to share yet. 19:10:25 ... So first we get evidence of wide review then we define the exit criteria, which could be passing tests in a 19:10:28 ... test suite? 19:10:29 nigel: Yes. 19:12:53 Courtney: Going back to the inline styling question, because I'm interested in getting to a published version 19:12:56 ... as soon as possible. 19:13:13 nigel: Thinking about the existing CSSWG response, and other potential responses, in my opinion it's quite 19:13:29 ... high risk to omit inline styling and require styling from a host HTML page. 19:13:57 Loretta: I support the desire to get to Rec quickly. 19:14:21 ... Previously the issue with inline styling was about syntax - the CG didn't agree it. 19:14:35 Courtney: I think the action there is to contact David and Silvia and get a status on that. 19:15:32 pal: Whatever decision they make, if there's a formal comment made then this group has to approve the response 19:16:36 ... and the commenter's response to that may be positive or negative. 19:17:41 tai: So the WG has to approve the edit and that the comment has been addressed. Do we have the expertise 19:17:49 ... on the spec to state that? 19:19:04 nigel: My understanding is that we can't proceed without the expertise but that we have enough members 19:19:22 ... from the CG like Silvia, Dave, Courtney, Loretta and Philip that we do have enough expertise. 19:19:49 Courtney: So we have a 4 step process: 19:20:04 ... 1. Get feedback. 19:20:12 ... 2. Someone from the CG writes a proposed response. 19:20:22 ... 3. The WG reviews the response and maybe edits. 19:20:29 ... 4. Get the response form the original commenter. 19:22:06 nigel: Step 2 is interesting - ideally I'd prefer for an individual who is a member of both the CG and WG and 19:22:45 ... can be representative of the CG, so that at least we have some WG backing for the responses immediately. 19:23:06 Courtney: I agree with that. I'm happy to talk to Dave about that - he may already have someone in mind to do that. 19:24:20 nigel: I want to record the actions - can I just check it makes sense to put them in the WG tracker? They're 19:24:22 ... WG actions. 19:25:04 tai: Makes sense. 19:25:13 Courtney: We're at a hand-over point between CG and WG. 19:25:51 tmichel: Does it make sense to use the dispositions tracker for this? 19:25:59 nigel: I would defer that until we know who is doing it. 19:26:13 Loretta: have other groups used different tools? 19:26:30 tmichel: You can do whatever you want. You can produce an Excel document or an HTML page that describes it 19:26:54 ... You do need links to the archive for reference. The dispositions tool tracks the process steps. 19:27:03 Loretta: In practice does everyone use it? 19:27:06 tmichel: no. 19:27:19 glenn: people use github and a variety of tools. 19:27:41 nigel: Let's record the actions. 19:28:31 Action: Courtney Determine who is responsible for putting together the dispositions of comments document and get a schedule from them. 19:28:31 Created ACTION-394 - Determine who is responsible for putting together the dispositions of comments document and get a schedule from them. [on Courtney Kennedy - due 2015-04-17]. 19:29:06 Action: tmichel Chase the accessibility WG for a date for getting WebVTT feedback 19:29:06 Created ACTION-395 - Chase the accessibility wg for a date for getting webvtt feedback [on Thierry Michel - due 2015-04-17]. 19:31:20 Courtney: if our response to comments is to change the spec do we have to redo the request for wide review? 19:31:31 tmichel: If you publish a new WD then you only need to ask for comments on the delta. 19:32:12 tmichel: If we publish a new WD for wide review we should send an email to the groups on the charter and 19:32:25 ... make sure it's archived. If Dave did that before then the emails are not on the archive. 19:34:02 nigel: For IMSC 1 we sent to all the charter dependency groups and the W3C external liaisons filtered by 19:34:12 ... interest area, BCCing the member-tt list so they were archived. 19:35:29 Action: dsinger produce evidence of request for wide review for WebVTT, for the archive 19:35:29 Created ACTION-396 - Produce evidence of request for wide review for webvtt, for the archive [on David Singer - due 2015-04-17]. 19:37:15 nigel: So we're deferring the decision on a second WD dependent on the disposition of comments. 19:38:00 nigel: Are the region semantics in WebVTT stable? 19:38:17 Loretta: I think that's at risk. Philip doesn't like the way it's specced currently. AFAIK it's only implemented in 19:38:46 ... Apple's players. It's not in Chrome, IE, Firefox or Opera. 19:38:53 Courtney: Why is it at risk? 19:39:59 nigel: I'm asking because I couldn't identify a resolution or consensus in the CG about the syntax and semantics. 19:40:14 tmichel: The WG has published it so it is a consensus in that sense. 19:40:29 Loretta: If it's not implemented widely then that won't do us any favours. 19:41:37 tmichel: What about the ::past and ::future pseudo-classes? I couldn't see support for that on any browser. 19:42:09 tai: From reading the spec I'm unsure in general how much CSS feature support is required. 19:43:23 nigel: While we're listing things we're curious about, what about the red box in §6.1 of the FPWD, which 19:43:33 ... looks incomplete? We can't go to CR with that there, right? 19:44:47 Courtney: It's not like that in the latest ED. 19:44:56 nigel: In that case there have been changes. 19:45:37 Loretta: There will be an action if we want to take a new snapshot. Someone needs to do a diff. 19:45:42 nigel: That's an Editor's job. 19:47:24 Loretta: Someone needs to understand what has changed in the ED to establish the diffs. The changes may 19:47:27 ... be unrelated to comments. 19:48:57 nigel: I've filed bug #28464 https://www.w3.org/Bugs/Public/show_bug.cgi?id=28464 for §6.1 19:49:07 tmichel: Does WebVTT do ruby support? 19:49:20 Courtney: Yes it supports ruby but not to the extent that's in TTML2. 19:50:47 nigel: One of my key points is to make sure that there's alignment between new features in TTML and WebVTT. 19:51:05 ... If the WebVTT Ruby implementation blocks future extension to get to the support level of TTML2 then 19:51:08 ... we should file a bug. 19:51:13 tmichel: Who decides? 19:51:17 nigel: THat's a WG decision. 19:51:44 Courtney: We can use that as a criterion for evaluating new feature proposals. 19:51:47 nigel: +1 19:52:28 tmichel: Another thing that has been discussed is if the CSS feature list should be a whitelist or a blacklist. 19:52:43 ... Dave, Nigel and I had different views of what the current wording means, and ought to mean. 19:52:50 Courtney: Can CSS support profiles? 19:53:07 glenn: No it does not, and there's no standard way to tell if a feature is implemented. There are some heuristics 19:53:27 ... and javascript libraries like modernizr that can be used to mitigate those issues. 19:53:43 ... One of the things I do is to set a value to something other than its initial value and then read its computed 19:54:01 ... value. It's not a guarantee that the feature is implemented, but it's a good hint if you can't even get that far. 19:55:24 nigel: Let's adjourn for lunch! 19:55:29 group: Strong agreement! 19:55:33 rrsagent, generate minutes 19:55:33 I have made the request to generate http://www.w3.org/2015/04/10-tt-minutes.html nigel 21:38:42 dakim has joined #tt 21:41:01 glenn has joined #tt 21:41:48 courtney has joined #tt 21:44:44 Topic: WebVTT <--> TTML Mapping 21:46:07 Courtney: We don't have too much to discuss with the group yet, but want to let you know about our plan. 21:46:22 ... We discussed this in Geneva and I've been unable to make much progress since then. We want more 21:46:38 ... structure around this so we can get it done. Andreas and I are working on a timeline and an outline for 21:46:55 ... a document. We welcome help from anyone who wants to contribute text. Andreas and I will divide it up 21:47:04 ... otherwise but if anyone else wants to help that would be great. 21:47:17 nigel: What will that look like? 21:47:34 Courtney: We haven't finalised anything. But let's think about the approach. Glenn brought this up in 21:47:56 ... Geneva, that mapping from TTML to WebVTT directly isn't the way to go, instead to go from ISD to WebVTT. 21:48:08 ... What we thought we could do is to map from IMSC to WebVTT. 21:48:27 glenn: There is a spec for the ISD format in TTML2, but the mapping semantics need to be inferred from the 21:48:48 ... rest of the spec but it's not perfect yet. A tool I've been working on, open source, based on TTX (TT transformer) 21:49:10 ... will take an arbitrary TTML document and spit out a sequence of ISDs. So that's an examplar of what the 21:49:22 ... semantics should be, but they're not written in detail in the spec. 21:49:45 Courtney: If our goal was to write sample code we could use that but our task is to write a document. 21:50:02 glenn: What that tool does is to flatten out the timeline and create a document for each interval during which 21:50:22 ... no changes occur, and also flattens out the styles into the ISD documents. So a lot of the complexity in TTML 21:50:40 ... disappears in that domain, and then it's easier to do a translation into some target format. If you don't do that 21:51:06 ... but focus on IMSC you still have to deal with the complexity of the styling system. 21:51:21 Courtney: Does IMSC require sequential timelines? 21:51:37 pal: No, not unless you have progressivelyDecodable set. 21:51:51 Courtney: My feeling is that its easier to start with a subset of TTML and then extend it later. 21:52:13 tai: It makes sense to start with IMSC because it simplifies things, has a similar target usage, to WebVTT, and 21:52:36 ... we might even start with a first draft based on something even more restricted to support the most 21:52:55 ... commonly used features. Then we can add more as it is needed. From a formal perspective for an algorithm 21:53:14 ... I think it may be the way to go to use ISDs, but it makes sense in the document to tackle certain features 21:53:40 ... like position, timing, style, explain how it works in each spec, and then at least readers will understand the 21:54:01 ... context and how it works. Even if any algorithm specified isn't used at least readers will be able to make 21:54:27 ... their own. I think we made a good start in Geneva that is well minuted. 21:54:43 ... Another thing we started is to work on the test files that we have that need to be built up for WebVTT. All 21:54:59 ... these features can best be described using examples. The missing test files for WebVTT could also be 21:55:01 ... generated. 21:55:20 Courtney: I think we're going to do some more work on this next week (me and Andreas). In our conference 21:55:44 ... call in 2 weeks we should have an outline and a schedule that we can share and discuss. We could circulate 21:56:18 ... it on email too. 21:57:12 nigel: It might make sense to use Mercurial to store it and allow it to be shared, or some other versioning tool. 21:57:18 rrsagent, draft minutes 21:57:18 I have made the request to generate http://www.w3.org/2015/04/10-tt-minutes.html tmichel 21:57:25 tai: That's publicly visible right? 21:57:28 nigel: yes. 21:57:36 Courtney: It would be good to have versioning. 21:57:48 tai: I think it's important to do a first draft and then take it step by step. 21:58:12 Courtney: I wanted to know if anyone wanted to contribute additionally. 21:58:22 Loretta: Speak to me about this - I'll see what I can do. 21:58:52 Courtney: It will also be good for developing the test suite. 21:59:13 Loretta, tai, Courtney: discusses actual plans for getting together to work on this. 22:00:04 nigel: Are there any particular inputs that you need e.g. in terms of expertise, understanding etc. 22:00:26 Courtney: I think we may need to call on Glenn for TTML expertise and Silvia for WebVTT! 22:00:45 glenn: One of these days someone will ask me to output WebVTT from my tool, and then I'll have to learn it! 22:00:53 Courtney: Are you available to do that? 22:01:02 glenn: Yes. The tool is all Java and is open source. 22:01:18 ... There's another project called the TTV TImed Text Validator. TTX is the transformer layer on top of that. 22:01:33 ... So this new timed text presentation engine is a layer on top of TTX, so it validates the TTML input, 22:01:41 ... translates it into ISDs and then formats them. 22:02:26 tai: Is there a place where we collate implementations? 22:02:35 glenn: It would be a good idea to add that to the wiki. 22:02:41 nigel: Sounds good to me. 22:02:55 ... though we wouldn't necessarily endorse them. 22:03:09 courtney has joined #tt 22:04:18 tai: It's good to let people know about this activity on the mappings in case people are also working on it. 22:04:35 Loretta: Probably good to make some progress before raising expectations that we then discover we can't meet. 22:05:37 nigel: Have you looked at the idea of picking the features that are easiest to match, even if they're from TTML2 22:05:44 ... like the tts:position attribute? 22:05:53 tai: I think we should start with TTML1 because it's more useful. 22:06:02 Courtney: We can certainly look at some low hanging fruit. 22:07:01 nigel: An outcome of this work might be to identify subsets of the two specs that easily map to each other, 22:07:21 ... and also any small additions to one spec or the other that would make the mapping much easier. We should 22:08:08 ... capture those as outputs of the work if possible. 22:08:37 tai: That would help with a more general desire to align the specs. 22:08:47 nigel: Knowing where the semantic overlap lies would be useful. 22:09:29 glenn: Sure, where there are things like ruby, let's make sure they align semantically. 22:10:45 nigel: I'm just saying if observations like that are made then we should capture them. 22:11:19 tai: If there are possibilities of making mapping easier, like in TTML2 to map the semantics to CSS then in 22:11:25 ... the long run that will make it easier. 22:12:04 Loretta has joined #tt 22:14:00 nigel: Looking at our charter, there are other documents that we could be working on, like the Note on live, 22:14:13 ... and the mapping of 608/708 to WebVTT. Where's that last one up to? 22:14:20 Courtney: Silvia wrote a second draft of that. 22:14:54 nigel: Has Silvia done that in the CG - if it's in our charter I'd hope to publish it from here. 22:15:05 Courtney: It's on the CG at the moment. 22:15:24 https://dvcs.w3.org/hg/text-tracks/raw-file/default/608toVTT/608toVTT.html 22:16:04 nigel: I see there's quite a lot of work there - great! 22:16:13 Courtney: Yes, and it's been updated recently. 22:16:25 tai: She also proposed a mechanism for putting inline styles into VTT. 22:16:42 Loretta: I don't think that's in the WebVTT spec yet. Certainly the original syntax for inline styles was never 22:16:48 ... adopted into the VTT spec. 22:16:59 tai: So a new proposal might overwrite this? 22:17:15 Loretta: Sure if a new syntax is adopted then that document should be updated to include it. 22:17:56 nigel: Thanks for making progress on the WebVTT TTML mapping document - it'll be good to make some 22:17:59 ... progress there. 22:18:41 ... Do you have a longer term schedule in mind? 22:19:00 Courtney: We're hoping for something in time for the Sapporo face to face but it's too early to commit to a 22:19:09 ... schedule until we've started working on it. 22:19:27 tai: Also we might size the task to fit a particular timescale, at least for a first iteration. 22:19:59 ... One last comment: is there an idea to have a common telephone conference where the WebVTT and TTML 22:20:11 ... communities can come together? Would that be of use? 22:20:32 Courtney: I think, speaking for myself, in the past it's been tough for Silvia because she's in Australia. So 22:20:47 ... finding a good time could be difficult. A standing meeting might not have much support but we can 22:20:52 ... schedule specific reviews as we need them. 22:21:03 tai: That would be good to have. 22:21:28 Loretta: We're ideally located to make finding meeting times maximally difficult! 22:21:39 Courtney: Some working groups rotate it. 22:21:55 pal: My suggestion is make sure there's a really clear agenda if you want people to get up in the middle of the night. 22:23:10 nigel: Here in a meeting of 7 members and 1 observer, approximately half the members here today have an 22:23:17 ... interest in WebVTT, so we're already doing it. 22:23:25 Loretta: Of course not all the skills are here though! 22:24:17 Topic: Actions 22:24:20 action-379? 22:24:20 action-379 -- Nigel Megitt to Obtain png images for #linepadding and #multirowalign test cases. -- due 2015-03-19 -- PENDINGREVIEW 22:24:20 http://www.w3.org/AudioVideo/TT/tracker/actions/379 22:24:36 nigel: I did those PNGs but now they need fixing. 22:25:04 ... But that's captured in Action-391, so I'll close this one. 22:25:08 close action-379 22:25:08 Closed action-379. 22:25:19 action-381? 22:25:19 action-381 -- Pierre-Anthony Lemieux to Prepare new list of changes to imsc 1 since cr1. -- due 2015-03-26 -- PENDINGREVIEW 22:25:19 http://www.w3.org/AudioVideo/TT/tracker/actions/381 22:25:39 pal: The next step is to close the outstanding issue and then prepare CR2. 22:25:46 close action-381 22:25:46 Closed action-381. 22:26:39 nigel: I think we need a new date for the SOTD and a summary of differences since CR1. 22:27:23 Action: pal Prepare CR2 transition version (after closing one remaining issue on IMSC 1) 22:27:23 Created ACTION-397 - Prepare cr2 transition version (after closing one remaining issue on imsc 1) [on Pierre-Anthony Lemieux - due 2015-04-17]. 22:27:40 atai2 has joined #tt 22:28:05 nigel: We also have two overdue actions, action-365 and action-378 which we can't complete either of yet. 22:28:50 pal: That's right. Do we want to make the request to SMPTE as part of the wide review request on the next WD 22:28:53 ... of TTML2? 22:29:11 nigel: We could, I'm not sure. No particular reason to wait that I know of. 22:29:18 pal: Okay, well that's ongoing. 22:29:53 Topic: Issues 22:30:03 nigel: I don't believe we have any pending review issues - no we don't. 22:30:34 issue-382? 22:30:34 issue-382 -- Require a computed non-indefinite begin time -- open 22:30:34 http://www.w3.org/AudioVideo/TT/tracker/issues/382 22:30:50 pal: So just to confirm, if I list the TTML1 features, there's a #timing feature. 22:30:59 glenn: That's a very generic feature. 22:31:12 pal: Okay, so there's no granularity to build on. 22:31:32 glenn: By the way the process for coming up with the features was subjective, and we used it to drive our 22:31:54 ... tests. You're supposed to test features, but very few specs list features. So we took the approach of creating 22:32:02 ... more fine grained features. 22:32:27 Topic: Sapporo TPAC F2F 22:34:34 nigel: Looking at the survey results there's already enough backing for a meeting to say there's consensus 22:34:40 ... to hold a meeting in Sapporo. 22:35:22 ... There's a clear preference for the Monday and Tuesday and no objections to that, so it's clear that we 22:35:26 ... should prefer those days. 22:38:47 nigel: In terms of joint meetings, someone has suggested CSS-WG, and another suggestion is HTML WG. 22:39:16 glenn: This is going to be difficult unless we make it short and extremely focused, like a single specific question 22:39:23 ... to answer. Otherwise it will be a waste of time. 22:39:48 nigel: My question is, what do people want to do there? 22:40:24 glenn: Bert Bos has made comments on WebVTT. He might be interested in the line balancing algorithms. 22:40:34 pal: I mistakenly hit the button for HTML WG and am going to unclick it. 22:40:50 tai: We mentioned over lunch that working with HTML to understand how TTML TextTrackCues can work. 22:41:08 glenn: We have a spec I wrote called the TTML API Level 1 and Level 2 (two specs) that try to define some 22:41:25 ... things on top of the TextTrackCue interface that somewhat mirror the WebVTTCue. But there are lots of 22:41:46 ... issues getting browser vendors to implement them. Microsoft and Opera might have an interest but 22:42:11 ... Google, Apple and Mozilla don't seem to be interested. But my position is that it would be a pretty good 22:42:44 ... way to go to use Javascript to populate a set of HTMLCues. Then you could transform into HTML on the 22:43:08 ... server and feed it using Ajax and HXR to populate it, or have client side javascript that translates the TTML 22:43:23 ... file locally, though that would probably be more complicated and less efficient than doing it on the server. 22:43:40 ... That approach would allow either browser based HTML rendering or create a fragment containing SVG with 22:44:01 ... a pre-rendered version of the cue. It turns out that inside the implementation in Webkit and Blink the 22:44:26 ... same approach is used, to convert WebVTT to HTML and use that. 22:44:47 pal: If this group has a consensus view that would be a good reason to meet with the HTML WG. 22:45:11 glenn: The most likely answers are either 1) no or 2) write that as an extension spec and bring it to the HTML WG. 22:45:46 ... There are various reasons for doing that - it's easier not to touch the HTML5 spec. Creating extensions is 22:46:02 ... easier, so that's how I would approach it. If that's the likely response then we could organise our inputs to 22:46:07 ... feed their expectations. 22:46:37 pal: There'd be value in saying how to do it, though we don't have a volunteer for how to do that yet. 22:47:20 glenn: The whole area of text cues in HTML5 is problematic in terms of testing and interoperability because 22:48:14 ... they used a subjective process for getting to Rec. Those areas were marked as at risk during the CR process. 22:48:31 ... Instead of addressing that they just published anyway. 22:48:36 tai: But it's in the Rec? 22:48:52 glenn: It is, so its final. The question is what comes next. The formal way that the HTML WG has decreed is that 22:49:14 ... if you want to do something that doesn't fix 5.0 or add features for 6 then bring it as an extension. 22:50:06 nigel: How does that approach interact with the WebVTT TTML mapping work? 22:50:12 Loretta: It's orthogonal. 22:50:17 glenn: +1 22:50:39 ... You can say that the lack of a TTMLCue is a motivator for the mapping work, because you could use the 22:50:54 ... mapping as a playback mechanism. I think its worth proposing a more generic cue proposal to the group 22:51:11 ... at some point. 22:52:11 nigel: That sounds like a great idea; so far we have nobody wanting to volunteer to work on that. 22:52:26 ... Just to test the view of the group, do we have consensus that an HTMLCue would be a good way to do it? 22:52:29 tai: +1 22:52:31 pal: +1 22:53:00 glenn: The good thing is that with a generic cue type then you can use any input format as long as you have 22:53:54 ... javascript to create it. I already have a fairly complete proposal from Opera proposing a way to do this in 22:54:00 ... blink. 22:54:07 tai: But that's not public. 22:54:22 glenn: Eric Lindstrom who was at our meeting in Geneva wrote it. 22:54:34 pal: Could we bring that document to the group for review and take it to HTML WG? 22:54:47 glenn: It's blink specific so it would need to be made more generic. 22:55:02 ... Maybe we could ask Eric if he wants to take this forward. 22:56:40 tai: I can contact Eric if you give me an action. 22:56:56 Action: atai2 Contact Erik Lindstrom re HTMLCue 22:56:56 Created ACTION-398 - Contact erik lindstrom re htmlcue [on Andreas Tai - due 2015-04-17]. 22:57:10 courtney has joined #tt 22:57:25 glenn: Opera has membership of this group via Philip Jägenstedt. 22:58:50 nigel: We do have consensus here that this is a good way to go, i.e. to propose to HTML WG HTMLCue as an extension 23:04:48 nigel: That's everything from the survey results that I need for my response to the organisers. That doesn't 23:04:59 ... close the survey - any new or amended responses will still be taken into account. 23:05:36 ... But it does mean we will go ahead with a Monday and Tuesday meeting. 23:05:48 ... Is there anything that we want to cover in that meeting? 23:06:05 tai: We should reserve some time for the mapping, regardless of what state we're in. 23:06:30 pal: Hopefully we'll be done with IMSC 1 but if not we should address any blockers. We should also work on 23:06:49 ... a 1st draft of IMSC 2, that should exist at that time. The hope is that IMSC 2 is simpler as a document 23:06:59 ... than IMSC 1 because all the features will be a subset of TTML2. 23:07:15 glenn: You will probably want an informative section mapping IMSC 1 features into IMSC 2 if they're no longer 23:07:29 ... present in IMSC 2. I imagine you might want to say something about migration, continuing use of old content etc. 23:08:11 nigel: For TTML2 maybe we should be targeting implementation reports. 23:08:24 glenn: I definitely hope we'll be in CR by then. 23:08:46 pal: When will you be able to show implementations of your tool? 23:09:13 glenn: The TTPE presentation tool I've been writing should be available at end of June. Probably end of August 23:09:22 ... to finalise the arabic, hebrew and bidi additions to it. 23:09:31 pal: That will remove a huge risk to TTML2. 23:09:48 glenn: I'm hoping to be able to get a lot of the new features into it as well. So we can have at least one 23:10:06 ... system to pass tests. I'm also adding support to the TTML Validator to validate the new constructs. 23:10:23 ... So in a sense we could point to the Validator as one implementation and the presentation engine as the 23:10:53 ... other implementation even though one is built on the other. 23:12:01 nigel: There is a question the Director will ask about who wrote the implementations. In this case we could have 23:12:19 ... an answer which is that the spec and both implementations were written by the same individual. I don't think 23:12:26 ... that he'd be that happy with that! 23:13:05 tai: A quick question about what TTPE does: from the documentation I'm not clear what the outputs are? 23:13:39 glenn: [draws on the whiteboard] There are 3 layers: TTV, TTX and TTPE. Generic TTML files can feed into 23:13:59 ... TTV - TTML1 or TTML2. It validates every feature of TTML1 and i'm enhancing it as I go to support TTML2 23:14:38 ... features. TTX converts into ISDs in a sequence which consists of multiple instances of an ISD. 23:15:16 ... TTPE then does two things: layout and then rendering. It takes the ISDs and creates an area tree which 23:15:44 ... is very similar to what's in XSL-FO, or a box tree in CSS terms. It has things like block areas, line areas, 23:16:01 ... glyph areas, space areas that fill the lines up. The rendering process then turns that into the chosen output 23:16:35 ... format (specified on the command line) - one of XML for debug, SVG or PNG. I've also specced out renderers 23:16:49 dakim has joined #tt 23:16:58 ... for html and vtt that I and the other engineers working on this haven't built yet. They just take the area 23:17:16 ... tree as the input and then write out whatever is needed to create the same visual output as is specified. 23:18:02 ... It's very easy for SVG. TTV and TTX are both in the TTV project in github. TTPE is a separate project. 23:20:10 pal: When that implementation is available, that removes a huge risk for TTML2 because there's at least one 23:20:31 ... baselins implementation for features. So it's hard for me to imagine not being able to enter CR in August. 23:20:39 s/baselins/baseline 23:21:15 Loretta: It looks like there's a lot of input for the TTML to WebVTT mapping there. 23:21:39 glenn: The question is how to map from the area tree to WebVTT. It's hard to work out how to do the positioning in WebVTT. 23:22:00 ... Since WebVTT doesn't have pixel positioning I may have to use the regions feature to do the layout. 23:22:17 ... One of the things about the mapping is that you can lose information without losing everything. At the basic 23:22:44 ... level I could pull the text out of the area tree and blast it into a WebVTT cue. That would lose a lot of information! 23:23:08 ... It's a question how to include positioning properly. 23:23:30 Courtney: Another item for the F2F in Sapporo should be the WebVTT progress. 23:23:48 dakim: Do you think the ruby stuff is important for WebVTT v1? 23:24:04 Courtney: I think it's less important than inline styling, and also we need to figure out how to do it in terms 23:24:19 ... of the design. A lot of what we need is already in CSS so maybe if you do inline styling in CSS maybe you 23:24:34 ... get all the ruby functionality you need. 23:24:52 dakim: I've got a bug open with Philip that you can't annotate the rt tag. 23:25:13 Courtney: If that's a minor change to the spec then we could consider adding that. I want to balance the features 23:25:21 ... against getting to a spec as quickly as possible. 23:25:51 nigel: Obviously that's a consensus issue for the group. 23:25:59 Loretta: the call is when it's needed not whether it's needed. 23:26:10 rrsagent, generate minutes 23:26:10 I have made the request to generate http://www.w3.org/2015/04/10-tt-minutes.html nigel 23:37:11 courtney has joined #tt 23:39:18 Topic: TTML2 topics continued 23:39:30 q+ pal to ask about ttp:contentProfiles "any" 23:40:30 pal: I'd like to walk through the positioning of images in TTML2 also 23:40:40 s/q+ pal to/pal: I'd like to 23:42:07 Loretta has joined #tt 23:42:37 dakim: I'm wondering why people would use extent and origin on content elements to define inline regions instead of just ipd and bpd 23:44:43 nigel: For ttp:contentProfiles what sense does "any" make? 23:44:59 glenn: I see what you mean - I can't see why an author would claim conformance with at least one of but not 23:45:06 ... all of a set of content profiles. 23:46:42 issue-386? 23:46:42 issue-386 -- There's no apparent use case for the "any" term in ttp:contentProfiles -- raised 23:46:42 http://www.w3.org/AudioVideo/TT/tracker/issues/386 23:46:45 reopen issue-386 23:46:46 Re-opened issue-386. 23:47:08 nigel: The next point was also raised by pal: the positioning of images in TTML2 23:47:27 pal: How do you position images? 23:47:43 glenn: If it's a background image then use the background position properties. If it's a non-background image 23:48:10 ... then you can treat it like a block or inline area so the rules apply, like ipd and bpd. Extent and origin do 23:48:17 ... not apply to content elements - they create regions. 23:48:32 ... There are two ways to do it - you can put a background image on a region. That's intended to support 23:48:50 ... non-content images. For content images e.g. with rasterised text in them one should not use the 23:49:10 ... backgroundImage mechanism but instead use image like in html. Put it in a block context e.g. as a child 23:49:33 ... of a div or in a p. So if you put a span with an ipd and bpd then it's now an inline block with an image in it. 23:49:48 ... If you want to do fine positioning you have to use spans with ipd and bpd. 23:50:16 pal: So if I wanted an image 80% from the top and 20% from the left of the viewport. 23:50:32 glenn: I'd create a region and define its position using the position attribute or origin. Then I'd put a div in it 23:50:50 ... with a single image inside that div. That would be the easiest way to position it. 23:51:23 ... If that's the only content in the region it would position it. You can also use ipd and bpd. 23:51:33 nigel: Have we got scaling semantics for images if we use ipd and bpd? 23:52:06 glenn: I thought so - I don't believe there's an issue. 23:53:11 issue-387? 23:53:11 issue-387 -- Consider applying ipd or bpd to content images -- raised 23:53:11 http://www.w3.org/AudioVideo/TT/tracker/issues/387 23:53:17 reopen issue-387 23:53:17 Re-opened issue-387. 23:54:24 pal: When I read the document the simple option wasn't obvious. 23:55:00 glenn: By the way displayAlign and textAlign would apply to it. 23:55:14 ... I think you may also be permitted to put an image in a body without a div. 23:55:40 ... No, body has to have div in it, div can have embedded class like image or other divs or ps. 23:57:38 nigel: The last point to raise is that content elements can have style attributes extent, position, origin, ipd and bpd all applied 23:57:53 ... simultaneously. I think the semantics are clear but some readers of the documents will find it hard to 23:58:34 ... understand. It may be useful to add non-normative examples or text to explain this scenario. 23:58:57 glenn: That's a feature of languages as they get bigger. I have an issue already to add recommendation text 00:00:34 ... to say that style attributes should not be put on content elements where they do not apply and can not be 00:00:37 ... inherited from. 00:01:20 nigel: It might also be useful to include an example document or non-normative text somewhere to remove 00:01:28 ... or reduce the likelihood of confusion there. 00:03:43 nigel: The last question I have is where we're up to on the CSS semantics mappings? 00:04:01 glenn: I have a proposal out for mapping to html that will generate that, but I don't have the resource to do that 00:04:06 ... at the moment. 00:04:13 tai: Didn't you have some draft material for that? 00:05:41 glenn: jdsmith picked up some work from Sean re change proposal 5 but that needs further attention. 00:05:48 https://www.w3.org/wiki/TTML/changeProposal005 00:06:18 glenn: When we started this work there was an idea that someone could create a Javascript shim that would 00:06:32 ... generate HTML/CSS. 00:06:52 ... CP5 needs to be carefully reviewed because the time evaluation algorithm wouldn't work. It also attempts 00:07:14 ... to specify some algorithms on how to resolve styles. We already have some normative language on 00:07:27 ... style resolution in the spec. My plan was to take that information and also timing rules and wrote them 00:07:40 ... into the spec in a readable fashion in the context of an ISD sequence then it would eliminate talking about 00:07:51 ... style and timing resolution in the context of an HTML mapping. 00:09:24 nigel: I was expecting equivalence notes under the style attributes to say what the equivalent CSS semantics 00:10:10 ... are to the already stated XSL semantics. I think CP5 is something different and much more complex. 00:10:38 glenn: I'm not sure how useful that would be. 00:10:45 nigel: I think it would be extremely useful. 00:11:24 glenn: In §11.3.1.3 and §11.3.1.4, when we originally talked about a CSS mapping the intent was to define 00:11:42 ... two subsections of §11.3.1.4 one based on XSL-FO and the other on CSS. That means taking part of the 00:13:08 ... material from CP5 and formally doing it in a way that is congruous with the XSL-FO. 00:13:59 tai: I understood this activity to be as Glenn described, but even just additional notes as nigel described would 00:14:16 ... have a big impact on readability, especially if the CSS semantic is first. 00:14:29 glenn: I'd be happy to put the CSS equivalents first. 00:14:44 tai: People don't know how close the features already are to CSS so it would make the document much more 00:14:48 ... accessible. 00:15:22 glenn: It would be insufficient just to create an equivalence table for each style attribute. We'd also have to do an equivalent 00:15:31 ... algorithm in §11.3.1.4. 00:15:34 nigel: I agree with that. 00:15:47 glenn: Nobody has been pushing me to do this much! 00:16:55 glenn: I won't be confident about the algorithm until I've seen it in code though. I'd be happy to use CP5 00:16:59 ... as a starting point at least though. 00:18:05 nigel: I think that's all on this and everything I had on the queue to deal with. In terms of your spreadsheet 00:18:14 ... glenn did you have any remaining technical questions? 00:19:26 glenn: I don't think so - I have a pretty good idea on most of the remaining things and I will come to the group 00:20:02 ... if there are any other questions. For example on disparity if there's anything unclear in the issue. 00:20:21 glenn: I'd like a response from SMPTE to know that they'll be satisfied with what we've proposed. 00:21:07 nigel: pal has the action to draft that. Good to know that there's a pressing need for that. 00:21:18 glenn: I'd like to draft some of the members to do some work particularly on the examples and images of 00:22:10 ... examples. 00:22:47 ... I have about 25 entries that have to do with creating example source content or images or both. I could do 00:23:06 ... it but my time would be better spent creating technical solutions. There's also an annex Q about a CEA708 00:23:13 ... mapping that it would be useful to have someone look at. 00:23:37 pal: Just point to RP2052-11 which already does it. 00:23:45 glenn: Does that point to SMPTE additions to TTML? 00:23:58 pal: I doubt it - it doesn't use image. I'm just offering an easy way to do that. 00:24:08 glenn: I don't know if that's going to be sufficient - it needs to be reviewed. 00:24:22 pal: I don't think we need that section - if nobody volunteers to write it we can remove it. 00:25:31 glenn: We've added 6 metadata items preceded with cea708 and others preceded cea608. I have a note to 00:25:49 ... elaborate those with reference to 608 and 708. If someone could sign up for doing some of the related 00:25:52 ... work items that would be useful. 00:27:03 glenn: In the end with TTML1 where we did this everyone had a role bringing some spec text in. 00:28:14 nigel: Okay, I appreciate the request and think that's a really good idea. It's not necessarily a good moment 00:28:28 ... to start assigning tasks now, but let's think about that and how to take it forward. 00:28:41 glenn: I reserve the right to ensure consistency of language and edit contributions. 00:29:33 nigel: Okay, we have the spreadsheet available (see yesterday's minutes) - everyone have a look at that and if 00:29:44 ... you want to contribute something contact me or glenn and let's share out those tasks. 00:31:21 glenn: nigel you raised issues on appendix L. 00:31:40 nigel: I've done the work to guess the new values for the conformance column against requirements - it's in the issue. 00:32:39 glenn: In Appendix M the element derivation needs to be updated for the new properties. 00:33:05 ... It would be useful to have a list of elements and attributes to add or change on the tables in M. 00:34:15 ... By the way you could edit the source document itself if you're set up to run ant, and I could help you set 00:34:36 ... up your build environment so you can run the authoring tools. We have an old process for generating the 00:34:54 ... spec using an XSLT-based translation process with a DTD. The language we write the spec in is governed by 00:35:20 ... XMLSpec that came out of the XML WG. 00:35:28 ... There's a stylesheet too. 00:36:27 nigel: Okay, that's interesting - let me know if you're interested in setting that up. 00:37:02 Topic: Meeting ending! 00:37:13 nigel: Thanks to netflix for sponsoring this meeting, and to Glenn for buying lunch. 00:37:21 glenn: And thanks to Netflix for sponsoring that lunch. 00:38:17 nigel: I think we've had a really productive two days, so thanks everyone for attending and contributing. 00:38:26 tai: And thanks for chairing and scribing too! 00:38:28 nigel: Thanks! 00:40:07 nigel: We've done all our goals for this meeting and also handled issues or at least come up with action plans. 00:40:11 ... So well done everyone! 00:40:16 ... [adjourns meeting] 00:40:19 rrsagent, generate minutes 00:40:19 I have made the request to generate http://www.w3.org/2015/04/10-tt-minutes.html nigel 00:51:29 nigel has joined #tt 00:53:03 glenn has joined #tt 02:57:45 nigel has joined #tt 03:00:41 nigel has changed the topic to: TTWG: next meeting 2015-04-16 1400 UTC 04:37:39 nigel has joined #tt 04:39:03 nigel_ has joined #tt 04:41:10 nigel has joined #tt 04:41:37 glenn has joined #tt 04:41:58 nigel has joined #tt 04:59:30 glenn has joined #tt 05:11:50 nigel has joined #tt 06:17:55 glenn has joined #tt 07:02:10 glenn has joined #tt 15:02:27 glenn has joined #tt 22:34:26 nigel has joined #tt