[whatwg] <video> element proposal

[Benjamin Hawkes-Lewis]

Håkon Wium Lie wrote:

> I think we want to make video a first class citizen of the web.

Arguably it already is. It just hasn't been implemented in a
easy-to-author, UA-interoperable, or end-user accessible fashion.

> That means, IMHO, that there must be a simple way to add video to HTML
> pages. 

How is that incompatible with the use of SMIL?

It seems to me adding video to web pages is complicated because of:

1) Poor support for a wide variety of formats.

2) No guaranteed end-user capability to play, pause, rewind, etc. (so
that authors must build such interfaces themselves with JS).

3) Lack of guidance for transcripting/subtitling/captioning/audio
description (Joe Clark's working in this area) and poor implementations
of technology for such features.

4) Different browsers support different techniques for embedding content
in text/html (e.g. embed vs object, ActiveX for MIME handling, the Eolas
problems in IE).

As it stands, the <video> element proposal is interesting, but it
doesn't solve any of these problems for current or future UAs. If we
require HTML5 UAs to support Theora and to expose a minimal subset of
functionality, then we could conceivably fix 1) and 2) over the next few
years (IE/WMP being the major barriers here). Whether we can fix 4)
seems to be partly dependent on American courts.

Mandating support for SMIL should help with 3). It may be mandating
support for Ogg would also help, but there seems to be some doubt (e.g.
from a BBC researcher) as to whether Ogg is really suitable for
"multiple audio & video streams in the same file for things like audio
description and multi-language subtitles, and alternate audio tracks":

http://lists.xiph.org/pipermail/xiph-rtp/2007-January/000461.html

> I don't think one shoulr rely on other languages for this,

In reality, aren't you always relying on a language /of some sort/ (e.g.
a wrapper language for video)? It's not like you're trying to express
the video data within HTML.

> Part of the reason why we could to this so quickly is the work we have
> done on SVG.

IANAL and don't know Opera's licencing policy anyway, but if it makes
any difference there is a SMIL 2.1 player whose source code is available
for use by third parties under the LGPL:

http://www.cwi.nl/projects/Ambulant/FAQ.html

--
Benjamin Hawkes-Lewis