W3C home > Mailing lists > Public > public-media-fragment@w3.org > January 2009

Re: ISSUE-2: What is the mime type of a media fragment? What is its relation with its parent resource?

From: David Singer <singer@apple.com>
Date: Tue, 27 Jan 2009 15:30:29 +0100
Message-Id: <p06240843c5a4c9b9f8d3@[]>
To: RaphaŽl Troncy <Raphael.Troncy@cwi.nl>
Cc: Media Fragment <public-media-fragment@w3.org>

At 15:21  +0100 27/01/09, RaphaŽl Troncy wrote:
>Dear Dave,
>>I think this is a confused question.  When a fragment is specified on a URI,
>>the user-agent separates the fragment from the 
>>URI and requests the main part of the resource 
>>over HTTP, and then interprets the fragment 
>>locally.  The HTTP server, which gives the MIME 
>>type, is unaware of the fragment.
>True, the UA will separate the fragment from the 
>URI. We think of having smart UA that will also 
>take the fragment, and converts it into some 
>additional http headers when making the request 
>to the server. The http server will therefore in 
>this case be aware of the fragment requested, 
>and even can decide to serve just this fragment 
>... When serving the resource, the server writes 
>in the http header the type of the resource is 
>serving. What should it write when it is serving 
>a single key-frame of a video corresponding to a 
>media fragment request where a single time point 
>has been specified?

OK, that would be a subject for the HTTP 
extension that allowed this, wouldn't it?  I see 
a few possibilities:
a) the MIME type of the requested fragment is the 
same as that of the original resource;  yes, that 
might result in one-frame movies, and so on;
b) the UA can indicate what MIME type(s) it would like in response;
c) the server can decide;
d) the specification for the base MIME type could 
be revised [but that way we go mad revising RFCs].

>>Even if it gets the fragment in the URI GET, it 
>>is (I think) supposed to ignore it.  So the 
>>MIME type of a fragmented resource is the MIME 
>>type of the resource, isn't it?
>This is what I would think too. But I would like 
>to get the confirmation, since it is not crystal 
>clear (for me).
>Best regards.

Well, let's step into another realm.  Suppose 
that we extend HTTP to allow asking the server to 
extract a piece of a ZIP archive;  it makes some 
(but not much) sense for the server to re-zip 
just that piece.

Overall, I have to say that the simplest is still (a)...
David Singer
Multimedia Standards, Apple Inc.
Received on Tuesday, 27 January 2009 14:32:52 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:52:41 UTC