Re: WebVTT spec

On Mon, 18 Mar 2013, Silvia Pfeiffer wrote:
> 
> I'm going through your changes right now.
> 
> I've just stumbled across this paragraph in addTextTrack():
> 
>     <p>Initially, the <span>text track list of cues</span> is not
>     associated with any <span>rules
>     for updating the text track rendering</span>. When a <span>text track
>     cue</span> is added to it,
>     the <span>text track list of cues</span> has its rules permanently set
>     accordingly.</p>
> 
> I suggest setting the default "rules for updating the text track rendering"
> to the WebVTT rules.

Why? Then you couldn't add a text track and add TTML cues to it (or 
whatever). That seems like a poor future design.


> This avoids uncertainty when creating a new TextTrack

What uncertainty?


> and it is backwards compatible. We may then need an extra optional
> parameter on addTextTRack() to change the assoicated rules, e.g.
> textTrack = media .
> addTextTrack<http://www.whatwg.org/specs/web-apps/current-work/#dom-media-addtexttrack>
> ( kind [, label [, language [, format ] ] )

Seems bad to require an argument that isn't necessary. It also introduces 
a new error condition (set the format to AML, then ad BML cues, and you'll 
get an exception; whereas in the current model it just works).

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

Received on Monday, 18 March 2013 19:34:47 UTC