W3C

Timed Text Working Group Teleconference

11 Feb 2016

See also: IRC log

Attendees

Present
Dae, Glenn, Mike, Nigel, Pierre, Shinjan, plh
Regrets
Andreas, Frans
Chair
Nigel
Scribe
nigel

Contents


<scribe> scribe: nigel

This Meeting

nigel: I propose to look at Action Items, IMSC issues and path to PR, TTML2, Profiles and Charter. AOB?

group: No AOB

Action Items

action-429?

<trackbot> action-429 -- Mike Dolan to Draft a wg note for the profile short name registry and ttml media type registration -- due 2015-10-08 -- OPEN

<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/429

nigel: I propose to close the action since work has begun, and just keep working on it.

mike: There are two things that need to be in the document, as I mentioned in my email. We some nearly published
... document to show to IANA to update the media type registration. Then the Note needs to be published so IANA can
... reference it.

nigel: In the past I thought we agreed that we need to put the registration into TTML2 - is that no longer needed?

mike: We need to do something that IANA can reference before TTML2 hits recommendation.

plh: That's fine to put the registration text into a Note.
... When the document is ready for review send an email to me asking me to do the rest.
... It's actually part of the W3C Registered Media Types.

<plh> https://www.w3.org/2002/06/registering-mediatype2014.html

<plh> http://www.iana.org/form/media-types

nigel: It would be helpful Mike if you could add issues to the github repo for the profiles document for any parts that are missing like media type registration.

mike: I'm going to copy and paste the TTML1 SE section only with the new attribute.

close action-429

<trackbot> Closed action-429.

<plh> current registration: http://www.iana.org/assignments/media-types/application/ttml+xml

Charter

nigel: I've made a few changes locally which I want to review with my AC rep before issuing a PR. In summary:
... Clarify in Scope that IMSC is to be a Recommendation (as opposed to just in the Deliverables)
... Add into Scope that the Group will publish a version of IMSC that is compatible with TTML 2 and will address the concerns of backward compatibility with prior versions of IMSC.
... I've also made some other editorial tweaks and removed the Milestones section, if it's not needed.
... Any views on if we need a Milestones section?

plh: It's not a common view that we no longer need a Milestones section. You can point to an external Milestones section.
... It's helpful for organisations to understand the duration of commitment by seeing the Milestones.

nigel: Is it fair to reference a Milestones section that's on the home page or the wiki?

plh: Yes it is. As long as the Charter points to where they are.

nigel: Okay, I'd propose to do that. Any other views on this or other aspects of the Charter?

plh: Sounds the right direction to me.

nigel: I've also updated the end date to March 2018, which seems reasonable.

plh: Yes, it's reasonable for you to do that, and that seems like a good target.

nigel: Andreas raised an email query if we can simply extend the current charter by 3 months?

plh: I would like to have a charter before the end of the month to take to W3M. If you can explain why you need longer
... then I could request an extension, but not otherwise.

dae: What are we doing with WebVTT?

plh: David Singer is proposing to keep WebVTT in the charter at the moment.

IMSC

pal: Thanks to Glenn and Nigel for making progress on outstanding issues on IMSC.
... I'm happy to report that there's a proposal for all issues not deferred to IMSC 2.
... There's one remaining issue that still is outstanding.
... The proposed resolutions on all those issues will not lead to substantive issues and are documented in pull requests.
... We're getting close to requesting transition to PR at month end.
... For today, can we look at pull req #154 designed to address issue #111.

https://github.com/w3c/imsc/pull/154

https://github.com/w3c/imsc/issues/111

pal: This was the result of a lot of work - I think all the people with an opinion on this have weighed in positively.

nigel: So we have no outstanding comments on the pull req?

pal: Yes. There are only +ve comments.
... The Pull Req has been outstanding for over a week. There was a typo fix 5 days ago.

nigel: Any objections to merging PR #154?

group: no objections

nigel: Okay, go ahead and merge Pierre.

pal: I'll do that after the meeting.
... On that topic, all the other PRs lead to no substantive changes, but some are pretty important, like an issue with the
... EBU-TT-D issues and examples.
... Can we go to issue #146?

https://github.com/w3c/imsc/issues/146

pal: This is the only open issue on IMSC 1 for which a PR does not exist.
... I did some archaeology on this and there was text added in 2013 whenever there was discussion of rectangles the
... words "including the boundary" were added. Glenn has some concerns here. I'd like to avoid unintended consequences
... but also be receptive to Glenn's concerns.

glenn: My point is that the term "boundary" is not defined. It's not a term of art that we have used previously in any
... TTML document. It's not a term that's used in CSS or XSL-FO, so it's undefined. Even if you follow the mathematical
... definition that is not adequate. All of the areas are closed areas mathematically, over R2 space, which would require
... you to deal with epsilon delta methods for defining whether a point in that R2 space is included or excluded from
... that area. It's not practical to that in practice. All computation that's done in the context of IMSC or TTML will be done
... at the pixel level. They can be conceptual pixels as opposed to what really appears on the screen but they need to be
... rasterised to a grid for the purpose of determining intersection. Secondly, there is no ambiguity once it is rasterised
... into a pixel grid whether a pixel is included or not because we're dealing with closed rectangles here.
... The term "border" which is a term of art in XSL-FO which corresponds to the term border box in CSS, is somewhat
... closer in the sense that it is that portion of the box geometry that includes the border if one is present. That is how
... I originally interpreted the intention of this language about the boundary but after discussion it does not seem that
... it is what was meant. My suggestion is to remove those parenthetical phrases "including its boundary" because as
... it's not defined it's currently useless.

pal: Would you be comfortable using the phrase "closed rectangle" instead?

glenn: I used the term "closed area" which a rectangle can be. I was speaking of the mathematical definition which you
... inferred might be used.
... I don't see the need to consider it at all. We might simply state somewhere that all region intersection shall be done
... in a pixel space.

pal: Imagine I define a rectangle from 0,0 to 100,100...

glenn: There's no way to define a rectangle that way in IMSC. You can only state an origin and an extent.
... That means that a rectangle or region whose origin is 0,0 and has extent 100,100 includes pixels 0-99 in both
... dimensions but excludes the point 100,100.

pal: Where does it say that?

glenn: It doesn't need to because it's obvious. If the width is 100 pixels then 0-99 is 100 pixels wide.
... If we're talking about real value pixels, let's say fractional, e.g. a %age that you start with and translate into pixels
... in that case we need a rounding rule. That would be the floor so in that case you handle that also.

pal: Where's that [floor rule] stated?

glenn: It's in code in implementations. I could research it but I think it's clear that it does not include pixel 100,100.

pal: I'm just trying to be unambiguous when we define intersections.

glenn: In my estimation by adding (including its boundary) you have made it ambiguous. Without that phrase there's a
... common understanding in all W3C technologies. I may be able to track down some definition but it's a bit unnecessary.

nigel: My view is we should remove the phrase - I can't see what harm can result.

https://www.w3.org/2015/Process-20150901/#substantive-change

pal: I'm happy to remove this wording as long as it doesn't cause someone to object to the transition.

nigel: I think we should just make the change and simplify the spec, and make it clearly not non-aligned with all other specs.

pal: In the list of changes should we consider this as subtantive or not substantive?

nigel: I think by the letter of the process it falls under category 3 in the process.

plh: It's possible for the Director to agree to move forward even with substantive changes. We're going to worry about
... Wide Review and the Patent policy.

glenn: The real question is if we think that this change will make a substantial change to how processing or testing occurs
... and if it will break something that currently works or vice versa, and I don't think that's true.

pal: I think we're all agreed that a slight change here should not require a CR. We're asking if we should note the change
... and make the case to the Director or not document the change. I'd be happy with either approach. I'd be happy just to
... list the change and make the case that in our opinion the change cannot have an impact on patent and are not
... cause for a CR.

nigel: I'd go with that.

pal: If the Director disagrees that they're substantive changes I guess we can just remove them.

<plh> I have an other call now

glenn: I'd note that we made semantic changes to TTML1SE relative to TTML1 that we decided were not semantic.
... It's up to us to decide what to do.
... I move that we accept the current pull requests that are open.

nigel: I'd like to confirm for myself that I've reviewed them all before going ahead with that.

pal: Can you do that today or tomorrow?

nigel: Yes

pal: I'm happy to merge them after that.

nigel: I'll add any comments to them including LGTM comments.

pal: By the way I like the idea of splitting out the examples and then we can include them using respec.js but FYI you get CORS problems on Chrome unless you relax local file access.

nigel: Great, I didn't know about respec.js being able to do that.

glenn: It only makes sense to do that for complete examples, and that's what we do in TTML too.

nigel: Agreed, those are the only ones I pulled out.

glenn: I don't see any other blockers to moving to PR.

pal: My hope is that we can transition to PR as soon as the CR exit date is reached.

glenn: Can we make a motion that subject to any further comments we'd like to proceed to PR so we can highlight it to other groups?

pal: I'd like to do that too.

nigel: Okay, we're out of time - apologies that we didn't quite get to everything on the agenda today. Meet same time next week. [adjourns meeting]

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.144 (CVS log)
$Date: 2016/02/11 16:18:07 $