[whatwg] Spec should require UAs to have control to mute/ pause audio/ video

[Ian Hickson]

On Thu, 7 May 2009, Bruce Lawson wrote:
> 
> I think the spec should explicity require UAs to provide a mehanism to 
> mute audio and to pause video, even if the controls attribute is not 
> set.

We can't require this. Imagine a user agent that is a jumbotron in Times 
Square, controlled by just a big scrollwheel and huge button in the middle 
of the road, and big speakers.

Why should such a user agent be required to provide a "mute" UI?


> This will be of great benefit to screenreader users, as well as to 
> people who work in shared environments and who may not be expecting 
> noise.

I agree that it would be good practice to have such an interface, but 
that's something for users to take up with their vendors, it's not 
something that the spec should get involved in.


> (There is a precedent for worrying about "mere annoyance": "User agents 
> should not provide a public API to cause videos to be shown full-screen. 
> A script, combined with a carefully crafted video file, could trick the 
> user into thinking a system-modal dialog had been shown, and prompt the 
> user for a password. There is also the danger of "mere" annoyance, with 
> pages launching full-screen videos when links are clicked or pages 
> navigated.")

This is more about security than "mere annoyance". The latter is just an 
afterthought here.


> Should autoplay be removed as an attribute of audio and video, precisely 
> to remove such annoyance from people who rely on aural browsing?

On the contrary, having this be a declarative attribute makes it easier to 
disable. If we remove the attribute, authors will do it in JS and we'll be 
powerless to remove it. Ironically, the attribute is present precisely 
because it can be annoying.


On Sat, 9 May 2009, Bruce Lawson wrote:
> 
> But that adds weight to my greater argument for a mandatory control on 
> autoplayed media allowing user to mute sound. Relying in the operating 
> system's volume control is not the option as that mutes the offending 
> sound *and* screenreader output, so the rest of the content is 
> inaudible.
>
> On Thu, 7 May 2009, Tab Atkins Jr. wrote:
> > I don't see why this should be a browser requirement.  UAs are
> > *allowed* to provide such a facility to their users, so if your UA
> > doesn't, complain until they do!
> 
> Perfect solution for those who live in Utopia. But as many users neither 
> choose their browsers, or are unable to change them, it's better to 
> require it of the browser.

The world in which the spec has any control over what browsers implement 
is far more Utopia than the world in which browser vendors respond to 
their users' needs, believe me. :-)

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'