W3C home > Mailing lists > Public > public-html@w3.org > March 2011

Re: Change Proposal for ISSUE-147

From: Ian Hickson <ian@hixie.ch>
Date: Sat, 19 Mar 2011 04:00:20 +0000 (UTC)
To: David Singer <singer@apple.com>
cc: Philip Jägenstedt <philipj@opera.com>, public-html@w3.org
Message-ID: <Pine.LNX.4.64.1103190347130.18930@ps20323.dreamhostps.com>
On Fri, 18 Mar 2011, David Singer wrote:
> 
> Please remember that there are sources that might not be seekable at 
> all.  For example, if I have a URL form to address a TV tuner, you are 
> either tuned in, playing at 1.0, or not.

Actually per spec you're allowed to start buffering, TiVo-style. You don't 
have to though, in which case the "earliest possible position" is 
continually equal to the "current playback position" and the media element 
spends the whole time on the brink of having "ended playback". It's a 
quality of implementation issue. The better UAs would let the user buffer 
indefinitely (subject to available resources, obviously).


> Similarly, a hypothetical URL that asks for the source to be your camera 
> cannot do anything.

(Not that hypothetical; the WHATWG spec has a proposal for this now, 
using blob: URLs.)

> If your connection is RTSP/RTP, you can ask for non-1.0 playback rates, 
> but the server might 'suck' and refuse.
> 
> So it might not be your browser (or mine) that sucks.

Well, "suck" would be a rather strong way to put it, but the browser could 
certainly be better if it didn't buffer in those scenarios.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
Received on Saturday, 19 March 2011 04:00:51 UTC

This archive was generated by hypermail 2.3.1 : Monday, 29 September 2014 09:39:23 UTC