- From: Silvia Pfeiffer <silviapfeiffer1@gmail.com>
- Date: Tue, 12 Mar 2013 09:10:16 +1100
- To: Ian Hickson <ian@hixie.ch>
- Cc: public-texttracks@w3.org
- Message-ID: <CAHp8n2kZve5EKKb6m8259wo+fs7k7CHNMf_K-9PvQ3BXQiA5dg@mail.gmail.com>
On Tue, Mar 12, 2013 at 4:45 AM, Ian Hickson <ian@hixie.ch> wrote: > On Sat, 9 Mar 2013, Silvia Pfeiffer wrote: > > > > There is the WebVTT file spec, which is used outside HTML as well, so > > this is separate. But then there are the objects in HTML that parse > > WebVTT. So, those need to be in some kind of "glue spec". Could be an > > appendix to WebVTT, I guess. > > I think that's the wrong way to look at it. It's the same kind of > reasoning that led to HTML4 and DOM2 HTML being separate specs, with all > the tons of bugs that created. > > Conceptually, WebVTT is an abstract language, like HTML, with a defined > data model and a defined set of rendering rules. It happens to have a > textual serialisation, and it happens to have a DOM API. Not all > implementations need to have both, not all users need to use both. > > However, they are both intimately related. We can't change one without > changing the other. Putting them in different specs, or making one a > second-class citizen, will just lead to spec bugs, and that will lead to > poor interoperability and unhappy authors. > I definitely don't want it to look like the DOM API is second-class. A different document structure is probably sufficient, or some more introductory paragraphs. WebVTTCue will be in the WebVTT spec then, right? Silvia.
Received on Monday, 11 March 2013 22:11:04 UTC