- From: Philip Jägenstedt <philipj@opera.com>
- Date: Thu, 24 Nov 2011 11:21:13 +0100
- To: "Thomas Steiner" <tomac@google.com>, "Davy Van Deursen" <davy.vandeursen@ugent.be>
- Cc: "Erik Mannens" <erik.mannens@ugent.be>, "Yunjia Li" <yl2@ecs.soton.ac.uk>, "Media Fragment" <public-media-fragment@w3.org>
On Thu, 24 Nov 2011 10:32:26 +0100, Davy Van Deursen <davy.vandeursen@ugent.be> wrote: > Hi Thomas, > > given [1], all the tests passes. However, as you indicate below, by lack > of low-level access to the media resources some test cases are not > supported in practice. Can you provide a list of these test cases (is > the list below complete or are there also other test cases that must be > marked as 'notImplemented')? I'll let Thomas answer in detail, but given the current state of browser implementations, I'd say the following cannot be supported via a polyfill: * stopping at accurate end time in #t=start,end * smpte and clock time formats * track dimension * id dimension. Or, in short, what *can* be supported via a polyfill is: * time dimension with start time in NPT format * spatial dimension As mentioned in the teleconf via IRC, I don't think it is acceptable to count polyfills as implementations for state transition purposes. That is not to say that they aren't useful, just that they aren't proof that a feature is implementable in shipping software with all the extra issues that come up there. -- Philip Jägenstedt Core Developer Opera Software
Received on Thursday, 24 November 2011 10:21:39 UTC