Re: A new proposal for how to deal with text track cues

Hi Sean,

On Fri, Jun 14, 2013 at 10:21 PM, Sean Hayes <Sean.Hayes@microsoft.com> wrote:
>>Be honest to yourself: how much further is TTML?
>>It doesn't have a rendering algorithm yet,
>
> Depends on what you mean by algorithm. It has a reference semantics for expected visual output. It's had that for 5 years or more;

OK, that's good enough for me to qualify as a "rendering algorithm",
though I don't think all potential combinations of features are
covered in the example image outputs.


> the fact that certain parties don't like the language its expressed in doesn't change that fact.  If you don't think that's a rendering algorithm then HTML and CSS have got along just fine without a rendering algorithm for decades, so we're good there.


Actually, the language in HTML and CSS has changed to accommodate that
need. That's why we now deal with actual algorithms in these specs. It
helps interoperability.


> Perhaps you mean that the re-expression of the visual semantics in HTML/CSS is not at rec status yet. That is true, however even the re-expression has been around for a couple of years now and has been implemented a couple of times.

Can you share a link to that? I'm not sure I've seen that.


>> It doesn't have a JavaScript API that browsers would need to implement it properly (so application developers can manipulate the format),
>
> Of course it does - in fact it had the API before the format even existed, it's XML so the XML DOM just works. If you mean we don't have a final published shim that layers it into the HTML5 track API, well no, we hardly could have because you keep moving the goalposts, but we are attempting to keep our draft in sync with your changes so as soon as you are done we will be too.

Excellent. :-)


>>it still has bugs that prevent interoperable implementations in details.
> If you know of such bugs please report them. The 70 outstanding issues are in fact extension requests for the next version based on implementation and roll out experience gathered over the last 5 years or so for 10's of thousands of hours of content There are multiple implementations that pass the test suite. It is however always the case that there will be implementations and users that don't follow the spec, and report that as interoperability issues that's just life and not a lot we can do as spec writers to change it
>
> WebVTT even has a validator:
> http://quuz.org/webvtt/.
>
> As does TTML:
> https://github.com/skynav/ttv
>
> Hope this helps bring you up to speed.

Indeed, thanks. :-)

Silvia.

Received on Saturday, 15 June 2013 10:39:57 UTC