Re: New approach to HTML Media Capture

On Tue, 19 Apr 2011, Dominique Hazael-Massieux wrote:
> 
> Back in December, Robin Berjon asked for feedback on the proposed way 
> forward for HTML Media Capture, a proposed mechanism to prompt the user 
> to record media from attached cameras and microphones instead of loading 
> from the file system: 
> http://lists.w3.org/Archives/Public/public-html/2010Dec/0064.html
> 
> We've published last week an update to the document that, instead of 
> using a "capture" parameter as part of the accept attribute value 
> specifies a new "capture" attribute for <input type="file">: 
> http://www.w3.org/TR/2011/WD-html-media-capture-20110414/

I still don't understand what problem this is solving.

This is all you need to indicate that the user should be able to use the 
camera (or the filesystem) to get data:

   <input type="file" accept="image/*">

This is all you need to indicate that the user should be able to use 
either the filesystem, camera, or microphone, to get any kind of file 
including an image or a video:

   <input type=file>

I also don't really understand the purpose of MediaFile/MediaFileData. It 
seems all the information that they expose except "codecs" and "bitrate" 
can be obtained directly from an HTMLVideoElement object already. We could 
expose "codecs" on HTMLVideoElement as well, except that it's not clear 
that any UA would actually be able to generate reliable results for this 
since you can only generate results for this if you already know the file 
format. Better, IMHO, to work on getting a single (or maybe two) common 
format. We could expose "bitrate" if there's a good use case for it; we 
probably will, in fact, in the future Metrics object.

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

Received on Tuesday, 19 April 2011 17:27:07 UTC