- From: Silvia Pfeiffer <silviapfeiffer1@gmail.com>
- Date: Sat, 15 Sep 2012 09:25:12 +1000
- To: Jean-Baptiste Kempf <jb@videolan.org>
- Cc: Philip Jägenstedt <philipj@opera.com>, public-texttracks <public-texttracks@w3.org>, David Singer <singer@apple.com>
On Thu, Sep 13, 2012 at 6:48 PM, Jean-Baptiste Kempf <jb@videolan.org> wrote: > On Thu, Sep 13, 2012 at 09:00:08AM +0200, Philip Jägenstedt wrote : >> I don't really see why we should try >> to accommodate for non-Web users, in particular until they >> themselves come to ask for it. > > And yet you took SRT, because it was very widely used by non-Web users, > in order to have an important traction and facility to convert them into > WebVTT. >From our discussions at VDD [1] I took away that non-Web tools are indeed keen for some header data in WebVTT, including kind (chapters/captions/subtitles/descriptions/metadata), language (ISO code), inline CSS styles (the limited allowed set for WebVTT), possibly even label (a human readable string of what the content is). I've even heard the request that we drop these attributes from the <track> element and and just rely on this data being inside the WebVTT file. This latter one we can't do IMO for several reasons: * not all formats provide this information (and <track> was developed to allow other formats, too - e.g. IE supports TTML). * we don't want to have to download and parse a potentially large list of file headers just to create a menu for the video element when the user most likely will only need one or two of them. Since we don't want to drop these attributes from <track> another alternative for non-Web use would be to create a separate meta-data file for a collection of WebVTT files and their video(s) that provides the resource description. Something similar to what we experimented with at Xiph with the ROE format [2]. Since this does not avoid the duplication of information and introduced an extra complication by having to deal with yet another file format, the discussion at VDD came to the conclusion that having the information duplicated inside the WebVTT file header for non-Web applications was the better solution. JB, would you agree? Cheers, Silvia. [1] http://www.videolan.org/videolan/events/vdd12/ [2] http://wiki.xiph.org/ROE
Received on Friday, 14 September 2012 23:26:02 UTC