Re: Timed tracks

On Fri, May 7, 2010 at 8:59 AM, Sean Hayes <Sean.Hayes@microsoft.com> wrote:
> Philip Jägenstedt>> "From what I have read of the TTML spec, I do not want to support it in Opera, mainly because it fails to make use of CSS for styling".
>
> I keep hearing that this is a big problem and I'm not understanding it, TTML does "make use of CSS for styling", it just happens to also make use of parts of XSL:FO and SMIL.

What is meant by that statement is "it fails to make use of *only* CSS
for styling".  Making use of CSS, plus XSL:FO and SMIL, still means
adding more code to the layout engine to handle the XSL:FO and SMIL
parts.


> Could you elaborate on exactly what it is you are looking for from CSS? And more particularly what version of CSS. TTML could use CSS selectors to apply its style properties, but that can't be it; because CSS selectors won't work on a text format like SRT, which you seem to think is a good direction.

The idea is to use the layout mode and values of CSS, so that no
additional work has to be done in the layout engine, and it will
automatically be possible to easily use future advancements in CSS on
timed text as well.  Ideally the syntax of CSS would be used as well,
so a new parser won't have to be created just for it.


>  If it's the generalised box model to support vertical text you are objecting to in TTML, then can we take it Opera will not be implementing CSS3 when it's done, as it is busily migrating those XSL modifications to CSS that TTML styling relies on. What advantage would you see in moving TTML back to a non-internationalised text model? Especially when an example of vertical text is included as a use case in http://wiki.whatwg.org/wiki/Use_cases_for_timed_tracks_rendered_over_video_by_the_UA

I'm not an expert at XSL:FO, so I can't say anything specifically, but
is XSL's "generalized box model" identical to how CSS specifies
handling vertical text?  If not, then it's a problem, since we'll have
two different layout engines for web content.

~TJ

Received on Friday, 7 May 2010 16:28:35 UTC