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

On Fri, Sep 28, 2012 at 9:58 AM, David Singer <singer@apple.com> wrote:
>
> On Sep 26, 2012, at 23:14 , Simon Pieters <simonp@opera.com> wrote:
>
>> On Thu, 27 Sep 2012 00:45:02 +0200, David Singer <singer@apple.com> wrote:
>>
>>>> If the use cases don't require blank lines, we can use a blank line to end a multi-line header.
>>>
>>> ouch.  can we use the blank line, as now, to end the headers?
>>
>> Why?
>>
>>> you seem to want to make it possible to mix headers and cues,
>>
>> No, I'm not interested in that at all.
>>
>>> which means we have to have a more complex way to tell them apart, and rules not to put headers after cues, and I am not sure at all why.  can you explain?
>>
>> As soon as you've seen a cue, the header part of the file is over. I don't see why that's harder than ending the header part when seeing a blank line.
>>
>> I think a blank line is more aesthetically pleasing and more in line with the rest of the WebVTT syntax (cues are separated by blank lines) than "." or "]]" or "##".
>>
>
> OK, so a blank line is easily defined, easily seen, and easily escaped.  "Something that looks like the first cue" is not.  Yes, it's definable, but why? I think it substantially less aesthetic to have rules about character sequences that have to be escaped, than about (blank) lines that have to be escaped.  Your proposal would make some character sequences invalid in any attribute-value; whereas terminating the header block on a blank line has no such problem.
>
> I am puzzled at both the complexity and the aesthetics here, honestly, and can't work out what's driving.

I am with Simon. We want to be able to cut and paste multi-line CSS
into the WebVTT header, including any blank lines that exist, without
having to think about it too much.

If we finish the header area of a WebVTT file not on a blank line, but
on the first valid cue, then we don't need to escape anything really,
because it is quite unlikely to have a "time --> time" pattern in
anything but a cue. We might want to escape "-->" if necessary, but
that's all.

If instead we finish the header on a blank line, then all the content
after that (up until the first valid cue) is parsed as invalid cues
and not as header. This is the current situation and it's
unsatisfactory in my opinion.

Escaping empty lines is something I'd really like to avoid - it puts
too much effort on the author and it requires reformatting CSS and
other multi-line header values too much.

Cheers,
Silvia.

Received on Friday, 28 September 2012 00:07:29 UTC