Re: TextTrackCue discussions

On Fri, Sep 20, 2013 at 6:36 PM, Jerry Smith (WINDOWS) <
jdsmith@microsoft.com> wrote:

> This is only true if we proliferate cue objects for specific caption
> formats.  There may be some expediency in doing that in terms of removing
> the attributes from the base object, but the core textTrackCue can parse
> different formats now.  This seems like a powerful benefit that should not
> be removed from it.  Has it been retained thus far?
>

I'm not sure what you mean by "the core [T]extTrackCue can parse different
formats now". I guess you mean that the original HTML5 CR [1] definition
can expose both text and HTML forms of different cue formats. That was true
in late 2012, but has not been true since Ian redefined TextTrackCue
without including text attribute or getCueAsHTML() operation.

Even so, we should distinguish between the ability to expose these forms
and the ability to parse the source cue format. Even exposing these forms
requires separate definitions of what is exposed based on the source cue
format; i.e., the text form and the HTML form will differ based on the
source cue format. So I don't know if we could ever say that what was
exposed was format independent. It may have been the case for
implementations that assumed the VTT model for exposing text and HTML
forms, but that was not required or assumed by the specification of these
interfaces, but only an implementation strategy in some UAs.

[1] http://www.w3.org/TR/2012/CR-html5-20121217/


>
> Jerry
>
> -----Original Message-----
> From: Silvia Pfeiffer [mailto:silviapfeiffer1@gmail.com]
> Sent: Monday, September 16, 2013 4:02 PM
> To: Jerry Smith (WINDOWS)
> Cc: Glenn Adams; Brendan Long; HTML WG (public-html@w3.org)
> Subject: Re: TextTrackCue discussions
>
> On Tue, Sep 17, 2013 at 5:09 AM, Jerry Smith (WINDOWS) <
> jdsmith@microsoft.com> wrote:
> > I understand that the existing attributes map to WebVTT semantics, and
> > to me that isn't desirable.  I'm not sure of the history that added
> those.
>
> For creating WebVTT cues in JS, it makes the most sense to have the cue
> settings as attributes of VTTCue, since they are not part of the cue text's
> markup.
> It's a good separation between cue text markup and cue box formatting and
> I don't see us changing that again.
> It's also now part of the TextTrack CG, so change proposals would need to
> be made there.
>
> Regards,
> Silvia.
>