search

w3c / ttml3

Timed Text Markup Language 3 (TTML3)
http://w3c.github.io/ttml3/
Other
6 stars 6 forks source link

Definition of tts:fillLineGap does not match itts:fillLineGap as specified in IMSC 1.0.1. #15

Open palemieux opened 6 years ago

palemieux commented 6 years ago

itts:fillLineGap as specified in IMSC 1.0.1 has been adopted by the industry.

tts:fillLineGap definition should be copied from IMSC 1.0.1, using the IMSC namespace, so that future versions of IMSC can reference TTML2.

skynavga commented 6 years ago

So, you want to create a dependency from core TTML specs to IMSC specs? Very ugly.

On Wed, Sep 27, 2017 at 10:27 PM, Pierre-Anthony Lemieux < notifications@github.com> wrote:

itts:fillLineGap as specified in IMSC 1.0.1 has been adopted by the industry https://www.dvb.org/resources/public/standards/a174_dvb_ttml_subtitling_systems.pdf .

tts:fillLineGap definition should be copied from IMSC 1.0.1, using the IMSC namespace, so that future versions of IMSC can reference TTML2.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/w3c/ttml2/issues/429, or mute the thread https://github.com/notifications/unsubscribe-auth/AAXCb6hXtHpzWPCnOenHFDgaeCSyaTF4ks5smyAfgaJpZM4PmtlP .

css-meeting-bot commented 6 years ago

The Working Group just discussed fillLineGap, and agreed to the following resolutions:

The full IRC log of that discussion <nigel> Topic: fillLineGap
<nigel> github-bot: https://github.com/w3c/ttml2/issues/429
<nigel> github: https://github.com/w3c/ttml2/issues/429
<nigel> Andreas: Thanks for raising this, I support it, and wanted to check it myself.
<nigel> .. As itts:fillLineGap is a complicated attribute, I see problems implementing it, and in
<nigel> .. compatibility, so it makes sense to keep the namespace.
<nigel> Pierre: We might do something subtler like reference IMSC 1.0.1 from TTML2, and then
<nigel> .. remove it from IMSCv.next. What's really bad is today the syntax and semantics don't
<nigel> .. match.
<nigel> s/syntax and/
<nigel> .. It's hard to understand if the TTML2 semantics are the same as the IMSC 1.0.1 semantics
<nigel> .. because the wording is so different.
<nigel> .. Either the wording is insufficient in IMSC 1.0.1 and should be fixed there or it should be
<nigel> .. made the same in TTML2. Implementers need to know that the behaviour is the same.
<nigel> Nigel: Okay, alignment is the issue here - we should define something with the same
<nigel> .. semantics in the same way.
<nigel> Pierre: I'd extend that to the namespace too.
<nigel> Andreas: I mentioned it before - there has been a bad experience changing namespaces
<nigel> .. going from the ttaf namespace to the ttml namespace, which broke things unnecessarily.
<nigel> Nigel: I can see the issue about namespaces - ideally the profile should be a pure profile
<nigel> .. of the spec, and not have this zigzagging up between the profile and the base spec.
<nigel> Pierre: Yes, in the future, maybe we just have the definition in TTML2 and remove it from
<nigel> .. IMSC, just referencing it by feature designator.
<nigel> SUMMARY: Majority view in favour of aligning semantics and syntax between IMSC 1.0.1 and TTML2, partly constrained by existing implementation work and adoption.
skynavga commented 6 years ago

I can probably support an approach that incorporates the syntax from IMSC and the semantics, to the extent that this is possible, but I will object to creating a normative reference from TTML2 to IMSC*. I will not object to an informative reference.

On Thu, Sep 28, 2017 at 8:55 AM, CSS Meeting Bot notifications@github.com wrote:

The Working Group just discussed fillLineGap, and agreed to the following resolutions:

  • SUMMARY: Majority view in favour of aligning semantics and syntax between IMSC 1.0.1 and TTML2, partly constrained by existing implementation work and adoption.

The full IRC log of that discussion Topic: fillLineGap

github-bot: https://github.com/w3c/ttml2/issues/429 github: https://github.com/w3c/ttml2/issues/429 Andreas: Thanks for raising this, I support it, and wanted to check it myself. .. As itts:fillLineGap is a complicated attribute, I see problems implementing it, and in .. compatibility, so it makes sense to keep the namespace. Pierre: We might do something subtler like reference IMSC 1.0.1 from TTML2, and then .. remove it from IMSCv.next. What's really bad is today the syntax and semantics don't .. match. s/syntax and/ .. It's hard to understand if the TTML2 semantics are the same as the IMSC 1.0.1 semantics .. because the wording is so different. .. Either the wording is insufficient in IMSC 1.0.1 and should be fixed there or it should be .. made the same in TTML2. Implementers need to know that the behaviour is the same. Nigel: Okay, alignment is the issue here - we should define something with the same .. semantics in the same way. Pierre: I'd extend that to the namespace too. Andreas: I mentioned it before - there has been a bad experience changing namespaces .. going from the ttaf namespace to the ttml namespace, which broke things unnecessarily. Nigel: I can see the issue about namespaces - ideally the profile should be a pure profile .. of the spec, and not have this zigzagging up between the profile and the base spec. Pierre: Yes, in the future, maybe we just have the definition in TTML2 and remove it from .. IMSC, just referencing it by feature designator. SUMMARY: Majority view in favour of aligning semantics and syntax between IMSC 1.0.1 and TTML2, partly constrained by existing implementation work and adoption. — You are receiving this because you commented. Reply to this email directly, view it on GitHub , or mute the thread .
palemieux commented 6 years ago

I can probably support an approach that incorporates the syntax from IMSC and the semantics, to the extent that this is possible

I would also use the IMSC namespace, so that IMSCvNEXT can simply reference TTML2.

skynavga commented 6 years ago

Yes, I presumed that would be required to incorporate the syntax.

On Thu, Sep 28, 2017 at 5:48 PM, Pierre-Anthony Lemieux < notifications@github.com> wrote:

I can probably support an approach that incorporates the syntax from IMSC and the semantics, to the extent that this is possible

I would also incorporate the IMSC namespace, so that IMSCvNEXT can simply reference TTML2.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/w3c/ttml2/issues/429#issuecomment-332993588, or mute the thread https://github.com/notifications/unsubscribe-auth/AAXCb53yJ0IkR6DFlA8naP06LjvdBIf3ks5snDBSgaJpZM4PmtlP .

andreastai commented 6 years ago

(I add this wide review comment to the existing issue to avoid duplication)

Syntax (including IMSC namespace), prose and examples of the tts:lineGap feature should be exactly as defined in IMSC 1.0.1. Keeping syntax, prose and examples identical signals to the industry that implementation of this feature is a long term investment.

The IMSC 1.0.1 specification text represents also the latest agreement of the group regarding this feature. The new text seem not to add any change of presentation processor behaviour. It is therefore not needed.

If it is not possible to keep syntax, prose and examples identical it may be best to remove this feature from the TTML2 draft. In this case the only reference to use it, will be IMSC 1.0.1 or an upcoming version of IMSC.

skynavga commented 6 years ago

I'm beginning to have second thoughts about making normative use of a non-TTML namespace in TTML2. If the argument here is we should adopt itts:fillLineGap instead of defining a TTML2 namespace form, tts:fillLineGap, then why wouldn't this same argument be used for

etc.

An IMSCvNext document that wants to be compatible with both TTML2 and IMSC1 could simply specify both flavors of the property. TTML2 processors will ignore the itts flavor and IMSC1 processors will ignore the tts flavor.

Wouldn't IMSCvNext documents already have to do something like this for all other mechanism for which TTML2 does not adopt the exact IMSC1 syntax?

palemieux commented 6 years ago

I do not think the same argument applies to all IMSC 1.0.1 extensions. For instance:

skynavga commented 6 years ago

On Fri, Oct 20, 2017 at 9:10 AM, Pierre-Anthony Lemieux < notifications@github.com> wrote:

I do not think the same argument applies to all IMSC 1.0.1 extensions. For instance:

  • in the case of smpte:backgroundImage, it might be that there are features of TTML2 image, e.g. content type signaling, that are desirable in IMSC 1.1 and not available in smpte:backgroundImage. In that case, smpte:backgroundImage would be deprecated and TTML2 image would be used instead in IMSC 1.1. This is the situation in the FPWD of IMSC 1.1.
  • in the case of itts:fillLineGap, TTML2 does not bring anything that itts:fillLineGap does not already offer, and itts:fillLineGap is already in use. In fact, the addition of tts:fillLineGap was motivated by its presence in IMSC 1.0.1.

No. We jointly agreed on tts:fillLineGap at a prior meeting with the understanding that in IMSC it would use itts.

  • In this case, TTML2 should match the IMSC 1.0.1 semantics exactly.

Is there a difference in semantics from your interpretation? If so, we should be able to readily fix that in TTML2 for tts:fillLineGap.

  • In addition, it is a gratuitous pain to ask users and developers and users to use a different XML namespace in TTML2.

It is significantly more gratuitous to require the addition of a non-TTML namespace to TTML2, as this will have serious impact on the spec, on TTML2 processors, and, further, introduces a very bad precedent.

palemieux commented 6 years ago

No. We jointly agreed on tts:fillLineGap at a prior meeting with the understanding that in IMSC it would use itts.

Please provide a pointer to such a decision.

Is there a difference in semantics from your interpretation? If so, we should be able to readily fix that in TTML2 for tts:fillLineGap.

The text is different, so divergence is possible. Please update the text to match IMSC 1.0.1, unless the latter is incorrect, in which case plase file an issue against IMSC 1.0.1.

t is significantly more gratuitous to require the addition of a non-TTML namespace to TTML2, as this will have serious impact on the spec, on TTML2 processors, and, further, introduces a very bad precedent.

There is simply no meaningful on the specification, which already declares at least 10 namespaces.

There is also no meaningful precedent since itts is a W3C namespace, under W3C control.

palemieux commented 6 years ago

Please provide a pointer to such a decision.

In fact, it looks like tts:fillLineGap was added by the TTML2 editor without a PR:

https://github.com/w3c/ttml2/commit/867f1af199b81258da6aa9d7586c305e2042accb

skynavga commented 6 years ago

On Fri, Oct 20, 2017 at 10:26 AM, Pierre-Anthony Lemieux < notifications@github.com> wrote:

No. We jointly agreed on tts:fillLineGap at a prior meeting with the understanding that in IMSC it would use itts.

Please provide a pointer to such a decision.

The actual property name fillLineGap was Nigel's proposal here [1]. Which came in a discussion at the London F2F where you said:

Pierre: Based on what I've heard the simplest thing is a boolean style attribute that signals ... the authorial intent, which we can add to IMSC1, then hopefully in the scope of IMSC2 ... and TTML2 then we can interface with the CSS folks and come up with something better.

and I pointed out (earlier in the discussion) the pre-existing discussion and issue recorded in TTML2 Issue #150, which was filed in March 2016, almost a year earlier, and then there was the even earlier TTML2 Issue

101, filed in November 2016, and based on the TTML2 tracker issue #302,

opened in March 2014.

https://www.w3.org/AudioVideo/TT/tracker/issues/302

Nigel: I think we need to be clear what the requirement is here. I think the simplest thing ... would be a boolean flag that says "do it as now" or "extend in the block progression direction ... the background so that the before edge of each glyph's background coincides with the line area's before edge ... and the after edge coincides with the line area's after edge".

https://github.com/w3c/ttml2/issues/150#issuecomment-192490492 > Is there a difference in semantics from your interpretation? If so, we > should be able to readily fix that in TTML2 for tts:fillLineGap. > > The text is different, so divergence is possible. Please update the text > to match IMSC 1.0.1, unless the latter is incorrect, in which case plase > file an issue against IMSC 1.0.1. > > t is significantly more gratuitous to require the addition of a non-TTML > namespace to TTML2, > as this will have serious impact on the spec, on TTML2 processors, and, > further, introduces a very bad precedent. > > There is simply no meaningful on the specification, which already declares > at least 10 namespaces. > This change would impact the definition of TT Style Namespaces (would require adding itts), and would also raise the issue of how to deal with other vocabulary defined in the itts namespace (or a future revision). > There is also no meaningful precedent since itts is a W3C namespace, under > W3C control. > The precedent is that, until now, with the exception of the use of core XML and XLINK NSs, all TTML related vocabulary is defined in a TTML namespace, i.e., is under http://www.w3.org/ns/ttml. Adding vocabulary outside this hierarchy will require significant changes in existing schemas and implementations and require the author to add a special case xmlns:itts declaration every time they use this one style attribute, a clear and unmotivated exception to what an author (and developer) would expect. > — > You are receiving this because you commented. > Reply to this email directly, view it on GitHub > , or mute > the thread > > . >
palemieux commented 6 years ago

The actual property name fillLineGap was Nigel's proposal here [1]. Which came in a discussion at the London F2F where you said:

None of these discussions covered namespaces.

Adding vocabulary outside this hierarchy will require significant changes in existing schemas

The only change is adding an xmlns:itts declaration. This is not an extensive change.

require the author to add a special case xmlns:itts declaration every time they use this one style attribute, a clear and unmotivated exception to what an author (and developer) would expect.

Authors are already accustomed to multiple namespaces in IMSC1, and to associating fillLineGap with the itts namespace.

and would also raise the issue of how to deal with other vocabulary defined in the itts namespace (or a future revision).

Hopefully there is no more of this going forward, assuming TTWG is responsive to adapting TTML2 to industry needs.

skynavga commented 6 years ago

On Fri, Oct 20, 2017 at 11:07 AM, Pierre-Anthony Lemieux < notifications@github.com> wrote:

Please provide a pointer to such a decision.

In fact, it looks like tts:fillLineGap was added by the TTML2 editor without a PR:

867f1af https://github.com/w3c/ttml2/commit/867f1af199b81258da6aa9d7586c305e2042accb

It was originally added as ttp:fillLineGap with a PR https://github.com/w3c/ttml2/pull/346. The change from ttp: to tts: was an editorial change (no PR needed) based on the resolution of https://github.com/w3c/imsc/issues/238.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/w3c/ttml2/issues/429#issuecomment-338266073, or mute the thread https://github.com/notifications/unsubscribe-auth/AAXCb-x5I-PeXVy-hlba80N0G912HxGYks5suNNhgaJpZM4PmtlP .

skynavga commented 6 years ago

On Fri, Oct 20, 2017 at 11:16 AM, Pierre-Anthony Lemieux < notifications@github.com> wrote:

The actual property name fillLineGap was Nigel's proposal here [1]. Which came in a discussion at the London F2F where you said:

None of these discussions covered namespaces.

Adding vocabulary outside this hierarchy will require significant changes in existing schemas

The only change is adding an xmlns:itts declaration. This is not an extensive change.

require the author to add a special case xmlns:itts declaration every time they use this one style attribute, a clear and unmotivated exception to what an author (and developer) would expect.

Authors are already accustomed to multiple namespaces in IMSC1, and to associating fillLineGap with the itts namespace.

and would also raise the issue of how to deal with other vocabulary defined in the itts namespace (or a future revision).

Hopefully there is no more of this going forward, assuming TTWG is responsive to adapting TTML2 to industry needs.

The industry is not of one mind on this particular point, so WG discussion is needed.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/w3c/ttml2/issues/429#issuecomment-338268193, or mute the thread https://github.com/notifications/unsubscribe-auth/AAXCbxcinpo05doUteEH3Cccw0822cUFks5suNVpgaJpZM4PmtlP .

nigelmegitt commented 6 years ago

I'd like to see a more extensive proposal for how to deal with the scenario like itts:fillLineGap if there is also a tts:fillLineGap that is otherwise identical (or indeed that is somewhat different).

Content providers:

Implementers:

Specifiers:

(those lists may not be complete!)

By the way, one option available to us is to define itts to be a TTML styling namespace. It's not pretty, but it is possible.

Nigel

From: Glenn Adams notifications@github.com<mailto:notifications@github.com> Reply-To: w3c/ttml2 reply@reply.github.com<mailto:reply@reply.github.com> Date: Friday, 20 October 2017 at 18:42 To: w3c/ttml2 ttml2@noreply.github.com<mailto:ttml2@noreply.github.com> Cc: Subscribed subscribed@noreply.github.com<mailto:subscribed@noreply.github.com> Subject: Re: [w3c/ttml2] Definition of tts:fillLineGap does not match itts:fillLineGap as specified in IMSC 1.0.1. (#429)

On Fri, Oct 20, 2017 at 11:16 AM, Pierre-Anthony Lemieux < notifications@github.commailto:notifications@github.com> wrote:

The actual property name fillLineGap was Nigel's proposal here [1]. Which came in a discussion at the London F2F where you said:

None of these discussions covered namespaces.

Adding vocabulary outside this hierarchy will require significant changes in existing schemas

The only change is adding an xmlns:itts declaration. This is not an extensive change.

require the author to add a special case xmlns:itts declaration every time they use this one style attribute, a clear and unmotivated exception to what an author (and developer) would expect.

Authors are already accustomed to multiple namespaces in IMSC1, and to associating fillLineGap with the itts namespace.

and would also raise the issue of how to deal with other vocabulary defined in the itts namespace (or a future revision).

Hopefully there is no more of this going forward, assuming TTWG is responsive to adapting TTML2 to industry needs.

The industry is not of one mind on this particular point, so WG discussion is needed.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/w3c/ttml2/issues/429#issuecomment-338268193, or mute the thread https://github.com/notifications/unsubscribe-auth/AAXCbxcinpo05doUteEH3Cccw0822cUFks5suNVpgaJpZM4PmtlP .

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://github.com/w3c/ttml2/issues/429#issuecomment-338275281, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AEwf8ycfwYIxfPWnh39tqDi6A_DWoeLYks5suNtzgaJpZM4PmtlP.

css-meeting-bot commented 6 years ago

The Working Group just discussed Definition of tts:fillLineGap does not match itts:fillLineGap as specified in IMSC 1.0.1. #429.

The full IRC log of that discussion <nigel> Topic: Definition of tts:fillLineGap does not match itts:fillLineGap as specified in IMSC 1.0.1. #429
<nigel> github: https://github.com/w3c/ttml2/issues/429
<nigel> Nigel: The last comment on this is from me, saying we need practical proposals to address the needs of
<nigel> .. the various constituents.
<nigel> Glenn: We already don't support, say, smpte:backgroundImage, so there's a precedent for
<nigel> .. not supporting fillLineGap seems no different.
<nigel> Nigel: That's confusing - not putting smpte:backgroundImage in TTML2 doesn't mean it
<nigel> .. cannot be in a profile.
<nigel> Glenn: That's right.
<nigel> David_Ronca: We've been advocating that IMSC vNext is a subset of TTML2, but there are
<nigel> .. two ways to get to it - either make sure that there's a TTML2 equivalent feature for everything,
<nigel> .. or by bringing IMSC namespace attributes into TTML2. We prefer the former.
<nigel> .. We'd like to address the features in TTML2 not by pulling in IMSC namespace.
<nigel> Nigel: Reminder that we agreed to look at each case individually last week, and that we
<nigel> .. should do that at TPAC.
<nigel> Glenn: Makes sense to me.
<nigel> David_Ronca: We should do that, at TPAC.
<nigel> Nigel: For fillLineGap specifically I'd like more detailed proposals.
<nigel> .. We have tts:fillLineGap in TTML2 and itts:fillLineGap and they have different wording, so
<nigel> .. they may have different semantics. The goal is probably to align the semantics, and then
<nigel> .. for IMSC vNext we need to think about how to handle this if both are permitted, for all
<nigel> .. constituents.
palemieux commented 6 years ago

As indicated in https://github.com/w3c/imsc/issues/276#issuecomment-343351123, tts:fillLineGap is removed from TTML2

skynavga commented 6 years ago

I believe the group did not finalize this conclusion, but at least one member required further consideration of the question (please correct me if I'm wrong).

On Fri, Nov 10, 2017 at 6:22 PM, Pierre-Anthony Lemieux < notifications@github.com> wrote:

As indicated in w3c/imsc#276 (comment) https://github.com/w3c/imsc/issues/276#issuecomment-343351123, tts:fillLineGap is removed from TTML2

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/w3c/ttml2/issues/429#issuecomment-343628424, or mute the thread https://github.com/notifications/unsubscribe-auth/AAXCbwigDiuc90XxL9qrMNtnw3bp15bGks5s1PbTgaJpZM4PmtlP .

css-meeting-bot commented 6 years ago

The Working Group just discussed Definition of tts:fillLineGap does not match itts:fillLineGap as specified in IMSC 1.0.1. ttml2#429, and agreed to the following resolutions:

The full IRC log of that discussion <nigel> Topic: Definition of tts:fillLineGap does not match itts:fillLineGap as specified in IMSC 1.0.1. ttml2#429
<nigel> github: https://github.com/w3c/ttml2/issues/429
<nigel> Cyril: We've discussed this today and agreed to remove `tts:fillLineGap`
<nigel> RESOLUTION: Following further discussion, group confirms that tts:fillLineGap is to be removed.