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

And having said that - maybe we can do something else for retrieving a
thumbnail than a fragment request. E.g. something like

http://www.example.com/movie.ogg?t=5m2s&type=text/png

This is a different resource to the movie file, so it can have a
different mime type.

Maybe we should consider taking advantage of the possibilities that
query and fragment identifiers both give us?

Cheers,
Silvia.


On Wed, Jan 28, 2009 at 9:45 AM, Silvia Pfeiffer
<silviapfeiffer1@gmail.com> wrote:
> I agree with Dave - the only problem where we are potentially changing
> mime types is the request for a picture out of a video. We should
> focus on this discussion. And.. no, I haven't got a good answer yet.
> :)
>
> Silvia.
>
>
> On Wed, Jan 28, 2009 at 1:48 AM, David Singer <singer@apple.com> wrote:
>>
>> At 14:36  +0000 27/01/09, Michael Hausenblas wrote:
>>>
>>> Dave,
>>>
>>>
>>>>  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;
>>>
>>> Sounds good. Didn't think about this one yet. But how do we technically do
>>> this? I fear I don't understand. Could you be more precisely on this
>>> option,
>>> please?
>>>
>>
>> Well, I am trying hard to think of a case *in multimedia* where the
>> statement
>> "the type of a piece of X *cannot* be the same as the type of X"
>> would be true.
>>
>> The obvious problem area is if you select a time-point in a video track of a
>> movie, then a fragment cast as a movie would have zero duration -- it's more
>> sensibly a picture.  Unfortunately, zero duration frames are explicitly
>> forbidden in MP4, 3GP etc. (since they can make the visual display at a
>> given time ambiguous).
>>
>> But this gets semantically tricky if there is sound;  what is the correct
>> representation of a point in time of a sound track?  It's not right to drop
>> it from the fragment (oof, we'd need media-type rules for what types get
>> dropped and what don't).
>>
>> This is steering me towards wondering if a piece of X, in time, necessarily
>> has some extension in time, i.e. a time-point is not a fragment (can you see
>> a zero-width character if you meet one in the street?).
>> --
>> David Singer
>> Multimedia Standards, Apple Inc.
>>
>>
>

Received on Tuesday, 27 January 2009 22:53:52 UTC