Re: The need of newlines in WebVTT (was Re: Displaying multiple lines in WebVTT)

On Wed, Apr 11, 2012 at 9:45 PM, Tim Hull <tim@globaltimoto.com> wrote:

> What do others think? Should we introduce an explicit line break
>> character such as<br>  and ignore any CR and LF characters inside the
>> cue?
>>
>
> It would be preferable if the form for breaking lines where an author
> expected them to occur could be automated, but I hazard a guess that it
> would be fairly difficult to gauge all the authors intended breaks.
>

Note that the word wrapping algorithm that you're talking about is (mostly)
an orthogonal question to \n vs <br>.

Once upon a time there was a word
> that often got into a lot of trouble because it kept getting lost.
>

The longer the longest line is, the more you'll have to move the eyes to
read the caption.  If you balance the lines above:

Once upon a time there was a word that often got
into a lot of trouble because it kept getting lost.

then you reduce the maximum length of any particular line, which means you
don't have to look around as much.  That's the algorithm I'm suggesting
currently, because it's a simple, easily-define approach that doesn't
involve heuristics, linguistic analysis or knowledge about particular
languages.

(I switched to a fixed-width font above, so we're not looking at different
proportional fonts.)

The wrapping algorithm could always be given as a guideline (rather than a
requirement), allowing UAs to experiment with "smarter" word wrapping
algorithms.  That wouldn't cause any new interop problems, since wrapping
isn't going to be deterministic anyway (it depends on things like fonts no
matter what you do).

-- 
Glenn Maynard