- From: Silvia Pfeiffer <silviapfeiffer1@gmail.com>
- Date: Thu, 4 Feb 2010 00:02:05 +1100
- To: Conrad Parker <conrad@metadecks.org>
- Cc: Media Fragment <public-media-fragment@w3.org>
I guess the only requirement is that the UA doesn't retrieve the full resource since that is counter-productive. This normally implies byte range requests. But I made it deliberately informative so what you say applies. Section 5.2.1 should probably make mote of what you're saying. Cheers, Silvia. Sent from my iPhone On 03/02/2010, at 11:39 PM, Conrad Parker <conrad@metadecks.org> wrote: > On 3 February 2010 20:57, Silvia Pfeiffer > <silviapfeiffer1@gmail.com> wrote: >> Hi all, >> >> We have both Firefox and Opera waiting for a "go-ahead" from us on >> what is ready for implementation and I don't really want to make them >> wait, since I believe we can do with a reality check. >> >> For this we should mark sections that we agree on as "ready to >> implement". But to make that call, we need a group agreement. >> >> I'd like us to move towards such a decision on the following >> sections: >> >> Section 5.2.1, http://www.w3.org/2008/WebVideo/Fragments/WD-media-fragments-spec/#processing-protocol-UA-mapped >> together with the specification part for NPT time: >> Section 4.2.1.1 >> http://www.w3.org/2008/WebVideo/Fragments/WD-media-fragments-spec/#npt-time >> and the general structure of media URI fragments. >> >> ... >> I believe what is required is to say basically: >> * this is the specification of such temporal URIs using fragments >> (normative) >> * when you get such a URI in a video element or in the URI bar, parse >> it in this way (normative) >> * then resolve it through byte range requests if you can (Firefox >> already is able to do this for Ogg though it takes multiple byte >> range >> requests to get to the right location; with new indexed Ogg files, >> this will improve) (informative) > > for this part (which only relates to how a client acts on a #t=), I > don't think it's necessary for the Media Fragments spec to say how a > user agent actually retrieves the data for a given media-type. This is > really dependent on that media-type; it may involve one or more > byte-range requests (eg. in the two schemes you mention for Ogg), or > it may involve loading auxiliary resources which describe where to get > various bits required for a time offset (eg. for quicktime adaptive > streaming or MS smooth streaming), and these may then involve requests > of many different resources that make up the media. Basically my point > is that the specific mechanism depends on both what the media-type > specifies and what the client is capable of, and doesn't belong in > this section of the Media Fragments spec. > > It might be simpler to just say that if the client is capable of > seeking in a resource, then the #t= indicates to the client that it > should seek to that time -- and perhaps we can even word that > unambiguously enough to be normative. > > Conrad.
Received on Wednesday, 3 February 2010 13:02:45 UTC