[whatwg] Timed tracks for <video>
jonas at sicking.cc
Fri Jul 23 00:05:28 PDT 2010
On Thu, Jul 22, 2010 at 11:31 PM, Ian Hickson <ian at hixie.ch> wrote:
> On Thu, 22 Jul 2010, Jonas Sicking wrote:
>> Based on your email, this sounds like a very interesting format.
>> However I can't really be certain due to two issues:
>> 1. Reviewing the format by reading the parsing specification is
>> basically impossible. Basically requires comming up with random
>> inputs, stepping through the specification for each input, until I see
>> patterns in the output. This will take more time than I have
>> available. It's also likely that I'll miss several patterns.
>> 2. In order to ensure that I've read all the parts related to WebSRT I
>> need to read the full HTML spec. While this is something I'm in the
>> process of doing, it'll probably take me months before I'm done.
>> In short, I'd love to see WebSRT be written as a separate spec, with
>> sections (non-normative if needed) that contain more human readable
>> information about the format.
> The WebSRT stuff itself is pretty modular, it's just two sections:
That is good to know, though it's unfortunate that this has to be
asked and isn't obvious by reading the spec. A related question is how
will I find out if this list changes. I think making these sections
available as a separate spec will significantly increase the number of
people that will review it.
Is there a reason not to, given that, as I understand it, there is
tool support for producing separate spec documents.
> The lack of introductory material and examples is absolutely a problem and
> I have every intention of adding much such material in the near future. I
> apologise for not having it already written; I generally delay writing the
> introductory material until after zcorpan, hsivonen, jgraham, Philip`,
> annevk, and others have looked at the spec, so that I don't have to
> rewrite the introductory material several times as it tracks the changes
> resulting from the major bugs they find. :-)
> In the meantime, the main place to start right now is the syntax of the
> WebSRT file, which is defined separate from the parser:
Good to know that this is on the way. I tried reading the section you
are quoting, but it was still more work than I have feel like I have
time for, especially given that I'm not an expert on timed text. I
suspect I'll just end up waiting until the introductory sections are
More information about the whatwg