W3C home > Mailing lists > Public > whatwg@whatwg.org > April 2010

[whatwg] Introduction of media accessibility features

From: Jonas Sicking <jonas@sicking.cc>
Date: Tue, 13 Apr 2010 10:35:47 -0700
Message-ID: <j2q63df84f1004131035i28fbb165z85a80e713895ec18@mail.gmail.com>
On Tue, Apr 13, 2010 at 7:50 AM, Eric Carlson <eric.carlson at apple.com> wrote:
>
> On Apr 13, 2010, at 12:28 AM, Jonas Sicking wrote:
>
>> Will implementations want to do the rendering of the subtitles off the
>> main thread? I believe many browsers are, or are planning to, render
>> the actual video graphics using a separate thread. If that is correct,
>> do we want to support rendering of the subtitles on a separate thread
>> too?
>>
>> Or is it enough to do the rendering on the main thread, but composit
>> using a separate thread?
>>
>> If rendering is expected to happen on a separate thread, then CSS is
>> possibly not the right solution as most CSS engines are
>> main-thread-only today.
>>
> ?It seems to me that the thread subtitles are composed, rendered, and/or composited on is an implementation detail that we should not try to spec. People are extremely sensitive to audio/video sync but we don't mandate how that should be handled, why start with captions?

I don't fully agree. For example we don't expose DOM Nodes to workers
because several browsers do not have thread safe DOM implementations.
Despite the fact that this is an implementation detail.

Similarly, while localStorage is easy to implement as specified on
single process implementations, it turns out to be a disaster on
multiprocess implementations such as IE8 and Chrome. If we had known
about this earlier we for sure would have adjusted localStorages API.
Despite the fact that this is due to an implementation detail.

I generally agree that implementation details should come second to
what makes a useful API. However if something is hard to impossible to
implement on certain classes of implementations, then we need to take
that into account. Especially when it seems like browsers are
generally moving towards those classes (in this case, multithreaded
audio/video).

That said, I'm not sure if we need to make such adjustments here. But
since I'm not sure I wanted to raise the issue.

/ Jonas
Received on Tuesday, 13 April 2010 10:35:47 UTC

This archive was generated by hypermail 2.4.0 : Wednesday, 22 January 2020 16:59:22 UTC