- From: Nigel Megitt <nigel.megitt@bbc.co.uk>
- Date: Thu, 16 Jan 2014 17:16:31 +0000
- To: Timed Text Working Group <public-tt@w3.org>
- Message-ID: <CEFDC465.179D9%nigel.megitt@bbc.co.uk>
Available at: http://www.w3.org/2014/01/16-tt-minutes.html
[1]W3C
[1] http://www.w3.org/
- DRAFT -
Timed Text Working Group Teleconference
16 Jan 2014
See also: [2]IRC log
[2] http://www.w3.org/2014/01/16-tt-irc
Attendees
Present
nigel, glenn, tmichel, plh, pal, jdsmith, frans_ebu
Regrets
Michael_Dolan_for_first_30-60_minutes,
Frans_de_Jong_after_first_30_minutes,
Pierre-Anthony_Lemieux_after_first_60_minutes
Chair
nigel
Scribe
nigel
Contents
* [3]Topics
1. [4]Charter
2. [5]Profile Discrepancy
3. [6]Individual character rotation
4. [7]Definition of "closest"
5. [8]IMSC language support - code points, xml:lang etc
6. [9]back to CLDR
7. [10]Pending Review Actions
8. [11]Actions overdue
9. [12]Change Proposals
* [13]Summary of Action Items
__________________________________________________________
<trackbot> Date: 16 January 2014
<scribe> scribeNick: nigel
nigel: will be taking part in chairs training next week,
choices are to proceed next week without me or to cancel and
reconvene on 30th.
glenn: suggests cancelling on 23rd
nigel: okay let's do that.
<scribe> agenda: Charter
Charter
nigel: some licence drafting has been going on in the
background, barring adverse comments expect to take new charter
to AC following edits by plh, towards the end of this week.
Profile Discrepancy
issue-162?
<trackbot> issue-162 -- update standalone profile documents --
open
<trackbot>
[14]http://www.w3.org/AudioVideo/TT/tracker/issues/162
[14] http://www.w3.org/AudioVideo/TT/tracker/issues/162
glenn: think we can close the issues.
... There was a question about backward compatibility. My
analysis showed that the previous versions were not up to date.
If anyone used them then those implementations need to change
anyway.
... They were not compatible with what was published in the
Rec. I'm happy with the new docs, which have been published.
nigel: Moved to Pending Review.
action-141?
<trackbot> action-141 -- Nigel Megitt to Call for Action to
identify discrepancies in TTML 1.0 on different profiles. --
due 2013-02-14 -- PENDINGREVIEW
<trackbot>
[15]http://www.w3.org/AudioVideo/TT/tracker/actions/141
[15] http://www.w3.org/AudioVideo/TT/tracker/actions/141
close action-141
<trackbot> Closed action-141.
action-225?
<trackbot> action-225 -- Glenn Adams to Work with plh to update
standalone profile docs for ttml1se -- due 2013-10-31 -- OPEN
<trackbot>
[16]http://www.w3.org/AudioVideo/TT/tracker/actions/225
[16] http://www.w3.org/AudioVideo/TT/tracker/actions/225
close action-225
<trackbot> Closed action-225.
Individual character rotation
issue-231?
<trackbot> issue-231 -- Individual character rotation --
pending review
<trackbot>
[17]http://www.w3.org/AudioVideo/TT/tracker/issues/231
[17] http://www.w3.org/AudioVideo/TT/tracker/issues/231
pal: The solution only affects horizontal glyphs with vertical
layout. Are there any other circumstances in which characters
may need to be rotated?
glenn: from the standpoint of normal writing, no, but for
arbitrary graphics it might be wanted of coure, for artistic
reasons.
... It's not supported now by CSS though it is supported by SVG
because SVG supports arbitrary angular rotation.
pal: it's possible to create custom fonts with rotated glyphs
so this is okay.
... I'm happy to close this.
close issue-231
<trackbot> Closed issue-231.
close action-255
<trackbot> Closed action-255.
Definition of "closest"
issue-259?
<trackbot> issue-259 -- Define "closest" -- pending review
<trackbot>
[18]http://www.w3.org/AudioVideo/TT/tracker/issues/259
[18] http://www.w3.org/AudioVideo/TT/tracker/issues/259
pal: See comments on issue. Wording would be better to define
in terms of distance from zero.
glenn: origin doesn't involve enclosure so rounding towards
zero seemed more appropriate than for other attributes where
making further from zero allows for more conservative spacing.
... We don't want to make a discrepancy between TTML2 and TTML1
by changing the way this could work.
pal: Reference to euclidean distance may mean that this never
gets implemented correctly.
... CFF-TT has specific wording for rounding from percentages
to pixels using 'round half up' terminology. So there's prior
implementation that does it differently, independently of
TTML1.
nigel: is this about the wording or the mathematics?
pal: no problem with the wording - it's about the mathematics.
glenn: there are 2 options in the issue, a and b. Is there any
difference in b between the proposal and the existing solution?
pal: I think so. Take tts:fontSize for example.
[19]http://www.w3.org/TR/ttml1/#style-attribute-fontSize
[19] http://www.w3.org/TR/ttml1/#style-attribute-fontSize
pal: Euclidean distance requires square root of the sum of the
squares. If two have the same computational result then choose
the one furthest from zero.
... My suggestion is to round each dimension separately.
glenn: this is a corner case. Prefer to think about, say,
padding.
[20]http://www.w3.org/TR/ttml1/#style-attribute-padding
[20] http://www.w3.org/TR/ttml1/#style-attribute-padding
glenn: the difference between the two things? For padding they
are the same, however distance is not otherwise defined. So I
used Euclidean distance specifically.
pal: Imagine there's a computed value for padding that consists
of 4 lengths - how would you compute the Euclidean distance
between this value and the next closest - what would be the
computation?
glenn: it would be on a component by component basis.
pal: I agree, so there's no difference.
glenn: the key is that this is already implied by the use of
the term "computed" which is already resolved to the component
level.
pal: Can we be clear about where "computed" means this?
glenn: The definition of CSS that we refer to defines this. We
presume that you include the CSS etc definitions.
pal: I think we need to be clear about how many components are
included in this Euclidean computation.
nigel: which values have multiple dimensions that are managed
together?
... is it only color?
pal: Euclidean distance can be applied to vector values as well
as scalar values.
glenn: In the case of padding it is not a vector. For fontSize
it would be considered a vector.
nigel: Aren't they two scalars?
glenn: In the case of scaling of bitmap fonts I definitely
meant to do a vector comparison.
nigel: That could be inferred from the notation used but do we
need to add wording to make it explicit?
glenn: As we've tried to become more precise it's become more
complicated. Do we change TTML2 to be different from TTML1, or
add an errata to TTML1? These are all informative notes at the
moment.
... We need to consider testing too.
pal: At least one reader did not have the same interpretation
as the editor.
glenn: 2 issues - one is using Euclidean distance; the other is
about origin in which you're suggesting changing the direction
of rounding.
pal: yes.
glenn: It may be worth separating these into different issues.
I'm more concerned about one than the other.
action pal to create a new issue to split these out for
tracking.
<trackbot> Created ACTION-257 - Create a new issue to split
these out for tracking. [on Pierre-Anthony Lemieux - due
2014-01-23].
glenn: if we put aside the issue of how distance is defined
linguistically then on this issue of origin, what are our
thoughts?
pal: I suggest deferring this issue until later and putting
higher priority on bigger issues.
... I can dig deeper into this and try to get some hard data
regarding the implementations prior to TTML1SE.
nigel: Okay we can take this offline, but in the absence of
further agreement we'll make no edits in TTML2
glenn: I'm happy to define what is meant by Euclidean distance
a bit, i.e. in terms of component-wise or vector-wise
comparison.
... In terms of rounding of origin I think it's early enough to
make a change now without impacting implementations now. We can
add an errata and a note.
... So we can solve both issues in a way that's not too complex
to implement.
<pal> sounds good
<glenn> ACTION: glenn to work with pal to resolve edits needed
for issue-259 comments [recorded in
[21]http://www.w3.org/2014/01/16-tt-minutes.html#action01]
<trackbot> Created ACTION-258 - Work with pal to resolve edits
needed for issue-259 comments [on Glenn Adams - due
2014-01-23].
IMSC language support - code points, xml:lang etc
action-238
<trackbot> action-238 -- Pierre-Anthony Lemieux to Follow up on
issue-295 -- due 2013-12-19 -- PENDINGREVIEW
<trackbot>
[22]http://www.w3.org/AudioVideo/TT/tracker/actions/238
[22] http://www.w3.org/AudioVideo/TT/tracker/actions/238
issue-295?
<trackbot> issue-295 -- Remove code point restrictions from
IMSC -- open
<trackbot>
[23]http://www.w3.org/AudioVideo/TT/tracker/issues/295
[23] http://www.w3.org/AudioVideo/TT/tracker/issues/295
pal: I have made the edits
nigel: Glenn are you happy with the updated wording?
pal: Annex B has been removed entirely and Annex A has been
refactored to take into account CLDR. The wording has been
changed to emphasise that this is a set of characters from
which
... authors are encouraged to choose depending on the language
of the element.
glenn: I still have some issues with table 2 because of its
non-inclusive nature. It is a somewhat Euro-centric subset of
languages.
pal: Remember this is a union of the CLDR sets, which still
apply - it's an addition to them only.
glenn: The scope is worldwide.
plh: I did my action item on ATSC - please review. We can also
close the action on the document licence.
... I have updated the charter for the document licence, which
only applies to WebVTT. I wanted to know where we are on
action-88 on the media type.
... If we don't conclude by Jan 30 there will be another long
cycle.
glenn: I haven't added the errata yet but given this info on
priority I'll pick it up this week.
plh: I did contact IANA to say we'll get back to them.
nigel: did you do the EBU wording?
<plh> ACTION: plh to add EBU wording in charter [recorded in
[24]http://www.w3.org/2014/01/16-tt-minutes.html#action02]
<trackbot> Created ACTION-259 - Add ebu wording in charter [on
Philippe Le Hégaret - due 2014-01-23].
plh: not yet.
pal: what's the process? Do we review it in the group before it
goes to AC?
<plh> [25]http://www.w3.org/2013/10/timed-text-charter.html
[25] http://www.w3.org/2013/10/timed-text-charter.html
plh: yes.
... I can do a diff too if that's helpful.
... I will send diffs to the public group
nigel: what's the review process?
plh: I would say a week for this group.
... regrets from me for the 30th though I'll be on IRC if
needed.
back to CLDR
pal: Am happy to follow up with Glenn and understand the
concerns.
glenn: My preference is to remove table 2 but otherwise to put
a paragraph in showing that it's a subset of example character
sets that are believed to apply to the specified set of
languages.
... To be exhaustive there would need to be citations for every
language but right now there are no references to normative
works.
... That paragraph should also say that others may be added in
the future.
pal: That sounds excellent.
glenn: I've noticed some issues elsewhere.
... for example lineHeight.
nigel: proposes to close actions 238 and 243 and leave 244
pending
close action-238
<trackbot> Closed action-238.
close action-243
<trackbot> Closed action-243.
pal: needs to leave soon. any more for me?
nigel: action-252 overdue
action-252
<trackbot> action-252 -- Pierre-Anthony Lemieux to Add an issue
for tracking purposes to refactor imsc to conform to the
content profile specification of ttml2 dependent on ttml2
profile updates. -- due 2013-12-19 -- OPEN
<trackbot>
[26]http://www.w3.org/AudioVideo/TT/tracker/actions/252
[26] http://www.w3.org/AudioVideo/TT/tracker/actions/252
pal: this is dependent on ttml2 work. I'll do it when I can.
nigel: will extend the due date so it's not overdue, to 7th
March.
... have done that.
Pending Review Actions
action-203?
<trackbot> action-203 -- Glenn Adams to Investigate 708 border
style mapping issues with mike's help -- due 2013-09-26 --
PENDINGREVIEW
<trackbot>
[27]http://www.w3.org/AudioVideo/TT/tracker/actions/203
[27] http://www.w3.org/AudioVideo/TT/tracker/actions/203
glenn: can't make progress with this as it stands due to lack
of definition.
action-214?
<trackbot> action-214 -- Glenn Adams to Review cp008 and
consider reformulating the recently added <animate> support to
better align -- due 2013-10-10 -- PENDINGREVIEW
<trackbot>
[28]http://www.w3.org/AudioVideo/TT/tracker/actions/214
[28] http://www.w3.org/AudioVideo/TT/tracker/actions/214
<jdsmith> I've just marked
[29]https://www.w3.org/AudioVideo/TT/tracker/actions/164
pending review as well.
[29] https://www.w3.org/AudioVideo/TT/tracker/actions/164
glenn: we discussed this in detail in the f2f and I've
implemented it. It has has a long review period.
<glenn> [30]http://www.w3.org/wiki/TTML/changeProposal008
[30] http://www.w3.org/wiki/TTML/changeProposal008
close action-214
<trackbot> Closed action-214.
action-219?
<trackbot> action-219 -- Glenn Adams to Change to rec in ttml2,
and add note to authors indicating that if author wants
specific value, then they should specify specific value -- due
2013-10-24 -- PENDINGREVIEW
<trackbot>
[31]http://www.w3.org/AudioVideo/TT/tracker/actions/219
[31] http://www.w3.org/AudioVideo/TT/tracker/actions/219
[32]https://dvcs.w3.org/hg/ttml/raw-file/default/ttml2/spec/ttm
l2.html#style-attribute-lineHeight
[32] https://dvcs.w3.org/hg/ttml/raw-file/default/ttml2/spec/ttml2.html#style-attribute-lineHeight
glenn: the action as stated has been done.
nigel: there's still a question of whether it's possible to
state a value as an alternative to the normal keyword's default
125% behaviour.
... The lineHeight algorithm is now more explicit.
glenn: I haven't described that algorithm previously on this
call.
... [takes a couple of minutes to describe the algorithm to
turn "normal" into a single computed parameter]
... I reviewed blink, webkit and mozilla as to how they resolve
"normal". They take the paragraph's nominal font, then look at
Ascent, Descent and line Gap, then take the font size that
applies and scale
... the metrics accordingly. They add the three values together
to give a computed lineHeight property when the original
specified value is "normal". If those metrics are not available
... then they use 125% of the fontSize that applied to that
paragraph.
... So at this level descendent spans with different font sizes
would not apply.
... and the 125% is only used if A, D and G are not available.
... There's a separate process for inline elements than for
block elements.
... The lineHeight becomes a minimum value for the line box
that will be used if there's nothing else in the line box that
has a greater height.
... However if there's an inline box that's a child of the
linebox that has larger non-replaced text then the A and D will
be used to increase the overall inline box's height.
... And there's separate processing for padding and margin.
... Note that when we introduce images we'll have replaced
content that comes into effect in this algorithm.
jdsmith: are the new steps consistent with those browsers?
glenn: yes, this is based on my analysis of those three. I
didn't analyse IE because I didn't have access to the source
code.
... Does MS use this algorithm?
jdsmith: Was there a different algorithm before?
glenn: We referred to CSS but that is ambiguous and defers to
implementations without documenting properly.
... Mozilla confirmed that this is what we meant but didn't
write down in CSS.
nigel: do we need to formally communicate this to CSSWG?
glenn: yes we do. In the past we've had comments at Last Calls
but should call this out specifically.
... The formal time to do it is during LC. I can raise it there
when we have consensus here on this language.
... I'd like to make sure that at least pal has an input
because he's written something different in IMSC.
action nigel to email group notifying of this change and
initiating formal approval
<trackbot> Created ACTION-260 - Email group notifying of this
change and initiating formal approval [on Nigel Megitt - due
2014-01-23].
action-221?
<trackbot> action-221 -- Glenn Adams to Double check spec
language on css and xsl-fo around line height stacking strategy
-- due 2013-10-24 -- PENDINGREVIEW
<trackbot>
[33]http://www.w3.org/AudioVideo/TT/tracker/actions/221
[33] http://www.w3.org/AudioVideo/TT/tracker/actions/221
nigel: this seems to be related to the similar edits.
action-260: reference changes in both action-219 and action-221
<trackbot> Notes added to action-260 Email group notifying of
this change and initiating formal approval.
close action-219
<trackbot> Closed action-219.
close action-221
<trackbot> Closed action-221.
action-233?
<trackbot> action-233 -- Glenn Adams to Reach out to CSS WG to
understand potential mapping of multi-row-align to CSS -- due
2013-11-18 -- PENDINGREVIEW
<trackbot>
[34]http://www.w3.org/AudioVideo/TT/tracker/actions/233
[34] http://www.w3.org/AudioVideo/TT/tracker/actions/233
glenn: I queried this, got the answer to use flexbox and
discovered that this is possible. I tested it on Chrome,
Firefox and Opera and Safari - it only worked on more recent
versions on Safari.
... I assume IE will work too because it supports flexbox too.
<glenn>
[35]http://lists.w3.org/Archives/Public/public-tt/2013Nov/0027.
html
[35] http://lists.w3.org/Archives/Public/public-tt/2013Nov/0027.html
glenn: Jerry you might want to have a look at the attachments
on this message including examples to try.
... I will also send a screenshot of what it should look like.
close action-233
<trackbot> Closed action-233.
action jerry and glenn to verify flexbox solution to multi row
align using IE
<trackbot> Created ACTION-261 - And glenn to verify flexbox
solution to multi row align using ie [on Jerry Smith - due
2014-01-23].
action-261: this is a follow-up to action-233
<trackbot> Notes added to action-261 And glenn to verify
flexbox solution to multi row align using ie.
glenn: There's still an open question as to how we specify this
in TTML2 - do we expose flexbox features directly or add some
specific line alignment syntax
... A new property is a shorthand for the use of the flexbox
solution.
action nigel to update relevant change proposal to include row
alignment requirement and potential CSS solution based on
action-233 findings
<trackbot> Created ACTION-262 - Update relevant change proposal
to include row alignment requirement and potential css solution
based on action-233 findings [on Nigel Megitt - due
2014-01-23].
Actions overdue
action-164?
<trackbot> action-164 -- Jerry Smith to Investigate interop
issues using CSS familyName constructs -- due 2013-12-05 --
PENDINGREVIEW
<trackbot>
[36]http://www.w3.org/AudioVideo/TT/tracker/actions/164
[36] http://www.w3.org/AudioVideo/TT/tracker/actions/164
jdsmith: actually this is pending review
... The view was that we should be aligning TTML with CSS.
There's a short term convenience for tools if we don't put the
CSS escape mechanism into the XML but it's cleaner to be
aligned and to
... use the CSS escapes.
... On another call we said to use XML escapes?
glenn: whatever tool we use the XML escapes to apply, because
we use XML. We don't use the stylesheets as syntactically
defined by CSS, just the value syntax.
... So the original use of these escapes came about because
when you express a CSS stylesheet as a separate resource then
you can use a backslash U notation.
... Since we don't reference external stylesheets that was not
an issue.
... In TTML2 we're using something closer to XSL-FO and SVG for
attribute value syntax in XML. If we introduce the escape
mechanism in tokens based on CSS's escaping mechanism then we
will end up
... with 2 mechanisms: the XML one and the CSS one. There's
redundancy there.
jdsmith: Yes, I prefer both.
glenn: This optimisation is in TTML1 as a constraint. In a
future version of TTML1 we could widen that out to re-include
the Unicode escapes.
... We can certainly include in TTML2 because strict backward
compatibility isn't required for this.
jdsmith: I will probably follow this up with a proposal for a
change.
glenn: Spec-ready text is preferred. The Change Proposal should
address interoperability with TTML1 re this issue as well, and
discuss the co-concurrent features of XML numeric character
references
... Should maybe define this in terms of where in the process
this happens.
Change Proposals
[37]http://www.w3.org/wiki/TTML/ChangeProposalIndex
[37] http://www.w3.org/wiki/TTML/ChangeProposalIndex
glenn: the 'index' doesn't actually have an index - it would be
good to edit this to make a true index.
action nigel to tidy change proposal index page to include an
index
<trackbot> Created ACTION-263 - Tidy change proposal index page
to include an index [on Nigel Megitt - due 2014-01-23].
action-263: glenn suggests listing a table with each change
proposal by number with different columns, with status, package
etc.
<trackbot> Notes added to action-263 Tidy change proposal index
page to include an index.
action nigel ensure all P1 and P2 issues not closed belong in a
change proposal
<trackbot> Created ACTION-264 - Ensure all p1 and p2 issues not
closed belong in a change proposal [on Nigel Megitt - due
2014-01-23].
Summary of Action Items
[NEW] ACTION: glenn to work with pal to resolve edits needed
for issue-259 comments [recorded in
[38]http://www.w3.org/2014/01/16-tt-minutes.html#action01]
[NEW] ACTION: plh to add EBU wording in charter [recorded in
[39]http://www.w3.org/2014/01/16-tt-minutes.html#action02]
[End of minutes]
__________________________________________________________
Minutes formatted by David Booth's [40]scribe.perl version
1.138 ([41]CVS log)
$Date: 2014-01-16 17:12:15 $
__________________________________________________________
[40] http://dev.w3.org/cvsweb/~checkout~/2002/scribe/scribedoc.htm
[41] http://dev.w3.org/cvsweb/2002/scribe/
Scribe.perl diagnostic output
[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.138 of Date: 2013-04-25 13:59:11
Check for newer version at [42]http://dev.w3.org/cvsweb/~checkout~/2002/
scribe/
[42] http://dev.w3.org/cvsweb/~checkout~/2002/scribe/
Guessing input format: RRSAgent_Text_Format (score 1.00)
Succeeded: s/authors/... authors/
Succeeded: s/gerry/glenn/
Found ScribeNick: nigel
Inferring Scribes: nigel
Default Present: nigel, +41.22.749.aaaa, tmichel, +1.720.897.aabb, glenn
, pal, [Microsoft], Plh
Present: nigel glenn tmichel plh pal jdsmith frans_ebu
Regrets: Michael_Dolan_for_first_30-60_minutes Frans_de_Jong_after_first
_30_minutes Pierre-Anthony_Lemieux_after_first_60_minutes
Found Date: 16 Jan 2014
Guessing minutes URL: [43]http://www.w3.org/2014/01/16-tt-minutes.html
People with action items: glenn plh
[43] http://www.w3.org/2014/01/16-tt-minutes.html
[End of [44]scribe.perl diagnostic output]
[44] http://dev.w3.org/cvsweb/~checkout~/2002/scribe/scribedoc.htm
----------------------------
http://www.bbc.co.uk
This e-mail (and any attachments) is confidential and may contain personal views which are not the views of the BBC unless specifically stated.
If you have received it in error, please delete it from your system.
Do not use, copy or disclose the information in any way nor act in reliance on it and notify the sender immediately.
Please note that the BBC monitors e-mails sent or received.
Further communication will signify your consent to this.
---------------------
Received on Thursday, 16 January 2014 17:17:02 UTC