[whatwg] Scripted querying of <video> capabilities

[timeless]

On Aug 14, 2008, at 11:14, timeless wrote:
> We'd probably be forced to lie and claim every codec imaginable.

[including ogg (or rather vorbis, theora, speex, ...), as these are
all "imaginable"]

On Thu, Aug 14, 2008 at 12:41 PM, Henri Sivonen <hsivonen at iki.fi> wrote:
> Would the situation be any different for the <source> element fallback?

yes. people wouldn't try to ask us questions we can't answer. instead
they'd be giving gecko lots of data and it'd figure it out.

> If MicroB.next ships without Gecko's built-in liboggplay <video> back end

no comment about plans for releases ;-)

> but ships with the GStreamer back end and a third party provides an Ogg
> plugin for GStreamer, as a user I'd want video on Wikipedia to Just Work if
> the third-party GStreamer Ogg plugin is installed.

yes, i'd expect it to just work. however i'd also expect the apis not
to work correctly. which means we'd probably be stuck with a case
where we either lie and say ogg isn't supported (because we have no
way to figure out if it's supported), which means it wouldn't work. or
we'd lie and say it is supported (which means that if someone sniffs
for ogg first, they'd send you ogg only and you would get nothing).

people should just provide <source> tags for all formats they have in
the order in which they'd like to send them. let the user agents
figure it out.

please don't encourage sniffing.

footnote: if someone's annoying/evil and only provides one <source>,
then yes, bad things will probably happen.

Could i put in a plea for browsers to consider flagging "this site
isn't nice, it probably won't work if you visit it with another
device, you should complain to the content provider". if possible,
useragents should be encouraged to encourage their users to demand
more availability of content in more formats :).