Re: Metadata in the VTT file header (bug 15851), use cases (and a need to close this)

On Sat, Sep 15, 2012 at 11:48 AM, Ian Hickson <ian@hixie.ch> wrote:
> On Fri, 14 Sep 2012, Ralph Giles wrote:
>> On 12-09-14 2:27 PM, Ian Hickson wrote:
>> >
>> > Wide usage was not a great concern. If it was, we'd likely have gone
>> > for something used a lot more, e.g. the formats used with TV
>> > broadcasts or on DVDs, which are used way more than SRT. SRT was
>> > mostly a niche format for video enthusiasts online.
>>
>> I...wow. I'm pretty sure the number of distinct entities creating
>> subtitles in SRT dwarfs those creating in DVD and broadcast TV formats.
>
> I don't have precise numbers, but I'd be shocked if the number of
> consumers of DVD subtitles or broadcast captions was lower than the number
> of consumers of SRT files. But I guess it depends what you're measuring.

I think you're speaking about two different things: Ralph about the
number of srt files in comparison to the number of DVD/broadcast
subtitles - and you about the number of consumers of these subtitles.
FWIW: I think you're both right. Not that it has any influence on the
outcome of this discussion.


In any case: I'm going to use file-wide metadata for the CEA-608/708
conversion document [1]. I'm going to follow what was last discussed
with Simon, except I don't really know when to end the multi-line
value, so I chose a random "##" as the character sequence to end it
(happy for any better suggestions). I preferred this as more obvious
than a single ".".

Single-line metadata example:

key: value

Multi-line metadata start with an empty key line "key:" and end with a
line that has only "##" as its content. The line-breaks (i) after the
key line and (ii) before the closing period are NOT part of the value.
The escape sequence for "##" could simply be "&#x23;#".

multi-line value example:

key:
valueline

valueline
…
##


I'm fully aware that the WebVTT spec itself needs not to say anything
about this - it will not break browsers in any case. It would,
however, be better if we could change the spec to regard the
multi-line values that may have empty lines in them not as broken
cues, but as part of the header. We could define - as Simon said -
anything between the "WEBVTT" magic string and the first successfully
parsed cue as "header".

Regards,
Silvia.


[1] http://dvcs.w3.org/hg/text-tracks/raw-file/default/608toVTT/608toVTT.html

Received on Monday, 24 September 2012 14:29:31 UTC