RE: {minutes} TTWG Meeting 2014-12-04

Re    <trackbot> issue-294 -- Style attribute to prevent overflow by shrinking text to fit on a line -- pending review

I am quite happy for this issue to be closed. I agree with Glenn’s assessment that this would be difficult to implement in the context of browser implementations.
My proposal was based upon some suggestions that CSS might be considering adding a rendering strategy where content adapted to layout, rather than vice versa.

For certain applications of text (e.g. subtitling), neither overflow or clipping are viable strategies when the rendered content exceeds the space available in the region.
However, avoidance of both overflow and clipping is an established practise in ‘broadcast’ subtitling (where font metrics are well known at authoring ☺


John Birch | Strategic Partnerships Manager | Screen
Main Line : +44 1473 831700 | Ext : 2208 | Direct Dial : +44 1473 834532
Mobile : +44 7919 558380 | Fax : +44 1473 830078<> |<> |

Visit us at
BVE, Excel London 24-26 February 2015 Stand No. N19

P Before printing, think about the environment

From: Glenn Adams []
Sent: 04 December 2014 16:40
To: Nigel Megitt
Subject: Re: {minutes} TTWG Meeting 2014-12-04

a couple of corrections/clarifications

On Thu, Dec 4, 2014 at 9:11 AM, Nigel Megitt <<>> wrote:
Thanks all for attending today's meeting. Minutes in HTML format can be found at:

TTML2 issues closure:

If there are no unresolved comments remaining we will close the following Pending Review Issues:

Thu 11th December TTWG meeting: issues numbered: 10, 230, 238, 239, 240, 241, 242, 244, 245, 246, 247, 248, 249, 250, 273, 287, 289, 290, 291 and 292.

Thu 18th December TTWG meeting: issues numbered: 21, 213, 236, 237, 285, 286 and 294.

Please review the solutions in the current editor's draft [TTML2ED] and raise any concerns offline prior to the meeting as far as possible, so that we can get through these in the time available.

Note that closure now will not prevent issues from being reopened or new related issues from being opened later for any reason, but allows us to verify that some apparently suitable solution has been implemented.



Minutes in text format:



                Timed Text Working Group Teleconference

04 Dec 2014

   See also: [2]IRC log




          glenn, pal, Thierry, Andreas, nigel, jdsmith, mike,









     * [3]Topics

         1. [4]This meeting

         2. [5]IMSC 1 transition to CR

         3. [6]Action Items

         4. [7]Issues

         5. [8]AOB

     * [9]Summary of Action Items


   <trackbot> Date: 04 December 2014

   <scribe> scribeNick: nigel

This meeting

   nigel: is there AOB?

   ... We should mention the ITU-R WP6B liaison

IMSC 1 transition to CR

   nigel: Pierre and I met the Director and plh and Thierry

   yesterday to discuss transition of IMSC 1 to CR.

   ... [result not minuted]


   <trackbot> action-333 -- Pierre-Anthony Lemieux to Create a one

   pager to cover the plan for the director's meeting for taking

   imsc1 to cr. -- due 2014-11-27 -- PENDINGREVIEW




   close action-333

   <trackbot> Closed action-333.

   tmichel: The document will be published on Dec 9 and then I'll

   send the call for implementation announcement

   ... on the same day. It will also be advertised on the W3C


   ... The webmaster has approved the document today.

   ... (so it is now frozen)

   nigel and pal: Thanks everyone for the work put into this


   glenn: Congratulations.

   pal: The next step is to start compiling our test suite and

   test material. I plan to translate the DECE

   ... submission into what I think is valid IMSC 1 documents -

   primarily a change to namespace.

   ... Also I'm going to reach out to implementors. Would the

   group also be interested in a more formal

   ... message or template message to send to those who are

   interested. How do we proceed?

   tmichel: The Call for Implementations is sent out to the chairs

   and ac-forum and will be published on the W3C homepage.

   ... That will be done on the day of publication. We can take

   that message and send it to whomever we want.

   ... We need to wait until publication.

   pal: Does that make it easy for people who want to contribute

   to know who to contact?

   tmichel: I'll send a link. Basically it says the spec is now in

   CR and invites participation for implementation.

   ... It also says how to feedback to the public mailing list.

   pal: Thank you. Also there are Simon's outstanding comments

   that have been deferred. Can we discuss that

   ... next week?

   nigel: Okay.

Action Items


   <trackbot> action-356 -- Nigel Megitt to Request review of

   ttml2 pending review issues from group -- due 2014-12-04 --





   close action-356

   <trackbot> Closed action-356.

   nigel: I should list the issues that have been implemented

   since last week, for review on 18th December,

   ... a similar timescale.

   glenn: The list is: 21, 213, 236, 237, 285, 286, 294. Those are

   all the new ones since last week.


   <trackbot> action-355 -- Glenn Adams to Resolve duplication

   between issue-357 and issue-229 -- due 2014-12-04 -- OPEN




   glenn: I've been focusing on the issues so I've not done work

   on this AI or any of the others.


   <trackbot> action-354 -- Mike Dolan to Investigate formal

   contact at arib -- due 2014-11-27 -- OPEN




   mike: Status is: I reached out to the executive director of

   ARIB and he's going to get back to me.

   ... Somewhat unrelated, a colleague from Toshiba indicated some

   other practical aspects of the ARIB work

   ... so that's interesting too. It looks like another 6 months

   to a year before it's out for basic UHDTV.

   ... Hopefully before the holidays we'll have specific logistics

   for how to work with them to add

   ... the extensions to TTML2.



   <trackbot> issue-21 -- window anchor points not supported? --

   pending review




   glenn: This is about positioning the root container region more

   flexibly, or other regions relative to it.

   ... I'm proposing a new tts:position style attribute fairly

   closely related to the backgroundPosition property

   ... in CSS 3 image backgrounds and borders. For example you

   could say "position the bottom edge of the

   ... region 20% above the bottom edge of the container region".

   ... So you could say 10% bottom, or center, which would center

   the region in its container region.

   ... Take a look at that - I think it will be interesting. I've

   defined it so that if you have both tts:origin and

   ... tts:position then in a TTML2 context you'd ignore

   tts:origin and use tts:position, otherwise you can

   ... still use tts:origin as a fallback. The presentation in

   that case may end up different of course.

   pal: Do we really need this feature? Positioning has been

   problematic for a number of implementors in

   ... the past so I'm concerned about the complexity. What is the

   use case?

   glenn: I'm responding to the issue - someone requested anchor

   points so I've provided a solution.

   pal: This issue was opened in 2008.

   glenn: We should bear in mind that there are general uses for

   TTML2 not just captioning and subtitling.

   ... There are features that need similar features. We shouldn't

   confuse implementation complexity with

   ... syntax complexity. Many browsers implement the CSS feature

   too, so there are implementations.

   mike: How would this work? If you have overflow, extent etc

   what would happen? Would the anchor remain

   ... where it was set. That needs to be addressed.

   glenn: This only addresses position of the region, not extent

   or overflow, which are separate.

   ... In order to compute the position you have to have resolved

   the extent already. The overflow issue

   ... is independent of positioning and sizing. So this is only

   related to position.

   mike: I think the combination of the two could result in some

   interesting effects depending on how it is

   ... defined.

   glenn: There are some extensions coming to the extent

   attribute, including the ability to say it should be

   ... computed so that the result is contained in the outer

   containing block, for example if you're sizing the

   ... root container region to ensure it fits in the available

   area of the related media object, and needs to be

   ... in the center, you'd specify extent="contain" and

   position="center" and the semantics would work

   ... alongside another new parameter storageAspectRatio which

   constrains the related media object's

   ... shape. Then the root container region is positioned


   mike: I think the original issue (could have been me, I'm not

   sure!) needs to be mapped from 708 semantics

   ... into this, whatever we end up doing. I share some of pal's

   concerns re complexity.

   glenn: I agree, and would point out that it's bad design in my

   opinion to throw out possible features early

   ... in the process. It's better to propose solutions that can

   be considered. As the process moves forward

   ... we can see if they will be implemented and do what is

   needed, rather than cutting them out too soon.

   ... This also establishes IPR precedent and commitment by

   putting them into the FPWD even if they don't

   ... end up in the final version. This is the wrong time to

   discuss complexity.

   atai: We discussed anchor points in the EBU group some years

   ago with Sean Hayes. We mostly discussed

   ... positioning of the region within the root container. We

   didn't talk about positioning the root container

   ... in relation to the media object. Just a note. I'm not sure

   if we have an option for simplification there.

   glenn: THere are three applications: 1) positioning the root

   container region wrt the display or the media object.

   ... For example a 16:9 display showing a 2:1 media - you may

   want to position the subtitles in the black

   ... bar beneath the video.

   ... 2) Positioning the region within the root container region.

   ... 3) Positioning background images relative to content or a

   region. The border rectangle, padding

   ... rectangle and content rectangle need to be considered.


   <trackbot> issue-213 -- Advance notice of deprecation for

   textOutline -- pending review




   glenn: There's no action - I don't propose that we don't

   deprecate textOutline.
I may have said this poorly, but I am proposing we DO NOT DEPRECATE tts:textOutline. In other words, I am proposing the NO ACTION option.

 I'll be drafting the

   ... orthogonal text shadow feature, but they're independent of

   each other.

   ... It's already in TTML1 so I propose to keep it.
"It's" = "tts:ttextOutline is"

   mike: As long as we end up having the 708 features that can't

   be emulated with textOutline (drop shadow support).

   glenn: The way we defined textOutline is that it extends

   perpendicular to the tangent of the outline of the

   ... glyph at any given point, inner or outer. textShadow is an

   x-y offset vector that translates the outline

   ... without changing its size.

   mike: My concern isn't textOutline but making sure we have text


   courtney: I think that means you won't be able to support all

   the FCC regulations in TTML.

   glenn: I'm proposing not deprecating, so both are available,

   and I believe it will support all of the options.

   ... We will need to go through all of the 708 edge styles and

   verify that they can be achieved.

   PROPOSAL: Close this issue.

   close issue-213

   <trackbot> Closed issue-213.


   <trackbot> issue-236 -- Character spacing, i.e. letter-spacing

   -- pending review




   glenn: this allows introducing tracking in the same way as CSS

   letter spacing. I've introduced it.


   <trackbot> issue-237 -- Inline space -- pending review




   glenn: This is similar - inline space allows the width of an

   inline box that contains some glyphs,

   ... for example a span with some content in it within a

   paragraph. I might want to say 14% of the region

   ... width, etc. It turns out that in CSS and XSL-FO there are

   both width and height properties that apply

   ... to content elements however the semantics are such that for

   inline non-replaced elements like span,

   ... width and height are ignored! So, what to do? It turns out

   there's a nice mechanism in CSS which I'd

   ... overlooked for a while, called display:inline-block. While

   you can't say that an inline block is a
s/inline block/inline area/

   ... particular width you can say that a block in an inline

   context is a particular width and height. So I

   ... introduced those concepts, and said that they magically

   turn spans into inline blocks to which width

   ... and height can apply. I've tested it in various browsers

   and it does exactly what you'd think it does.


   <trackbot> issue-285 -- Align rendered rows within a region to

   each other, and the set to the region -- pending review




   glenn: This is the long standing EBU request for multiRowAlign.

   We've played around with some options.

   ... We looked at flexbox, but I recently realised that the

   inline block mechanism I just described can

   ... satisfy the same requirement. E.g. A paragraph with a

   single span in it, the p being the whole width of

   ... the root container region, and you want to centre the text,

   but align any broken lines left or right to

   ... each other within that centre-aligned paragraph. The inline

   block mechanism allows that too.

   ... You can specify a separate alignment on the content of the

   span and that will allow two alignments,

   ... one to the content of the paragraph and the other to the

   span. My experiments on browsers show

   ... this works as desired. If you specify a textAlign on span,

   which is new, then it results in the promotion of

   ... the span to an inline block for display purposes. That

   would allow for example left justified lines in a

   ... center aligned paragraph.

   ... This separates out the concepts so the existing semantics

   don't have to change.

   atai: This is an interesting approach that the EBU group should

   review. I'm not sure it matches the intended

   ... presentation - does it align multiple lines of text with

   the longest line in the group? For example

   ... a centred first (and longest) line, with the second line

   right aligned to it? Would this work?

   glenn: I'm not sure. I need a written down sample or example to

   check. My proposal does work when

   ... the line breaks are defined with br elements. If two lines

   are wrapped in a span with textAlign="right"

   ... and contained in a paragraph with textAlign="center" then

   it would first layout the block right aligned

   ... and then centre the whole block as the paragraph. So this

   would have the result you want.

   atai: ok

   glenn: I'd be interested in scenarios where this might not

   work, according to my reading of the EBU spec.


   <trackbot> issue-286 -- Extend the background area behind

   rendered text to improve readability -- pending review




   glenn: I fixed that by allowing padding to be assigned to a

   span, and saying if done, it must use

   ... box-decoration-break:clone semantics.


   <trackbot> issue-294 -- Style attribute to prevent overflow by

   shrinking text to fit on a line -- pending review




   glenn: This is one that I believe John Birch specified and I

   said that the proposal was about content

   ... fitting, e.g. automatically reducing the font size to fit

   the content. Nobody in CSS land or elsewhere

   ... is doing that, so it would be highly complex to implement.

   However the region can now be shrink-fit

   ... to the content. I've proposed a new value for tts:extent:

   "fit-content" that comes out of a CSS3 spec.


   <trackbot> issue-307 -- Conformance language and processor

   profile rather than content profile. -- pending review




   nigel: Since relevant changes have been made I propose to close

   this with no further action.

   close issue-307

   <trackbot> Closed issue-307.


   nigel: Please note the liaison from ITU-R WP 6B which came in

   earlier today and went to member-tt.

   ... They would like a liaison I think, want to know about our

   work on WebVTT and IMSC 1 and how

   ... they relate to each other. They also have some detail

   questions on IMSC 1.

   pal: I'm happy to draft a first response on those IMSC 1


   <scribe> ACTION: pal Draft response to ITU-R liaison re IMSC 1

   questions. [recorded in


   <trackbot> Created ACTION-358 - Draft response to itu-r liaison

   re imsc 1 questions. [on Pierre-Anthony Lemieux - due


   glenn: I notice the liaison also mentions the ARIB-TT


   nigel: Yes, I encourage everyone to have a look at the liaison


   ... For next week we have 1 hour set aside and a lot of issues

   to close on TTML2, hopefully, so in the

   ... interests of time please could everyone discuss them

   offline on the reflector as much as possible, so

   ... we only have to discuss the minimum number of questions?

   [adjourns meeting]

Summary of Action Items

   [NEW] ACTION: pal Draft response to ITU-R liaison re IMSC 1

   questions. [recorded in


   [End of minutes]


    Minutes formatted by David Booth's [24]scribe.perl version

    1.140 ([25]CVS log)

    $Date: 2014-12-04 16:06:51 $



This message may contain confidential and/or privileged information. If you are not the intended recipient you must not use, copy, disclose or take any action based on this message or any information herein. If you have received this message in error, please advise the sender immediately by reply e-mail and delete this message. Thank you for your cooperation. Screen Subtitling Systems Ltd. Registered in England No. 2596832. Registered Office: The Old Rectory, Claydon Church Lane, Claydon, Ipswich, Suffolk, IP6 0EQ

Received on Thursday, 4 December 2014 17:34:46 UTC