[theora-dev] Re: Ogg Internet Drafts - create application/ogg-vorbis, application/ogg-tarkin, etc.

[Dan Miller]

> From: David Wheeler [mailto:dwheeler at ida.org]
... 
> 
> The #1 use of a MIME type is to launch an application to process
> the data.  If the MIME type doesn't include information on the
> container format, then the application won't know what to do with
> the result.

You might as well ask, how does an application know what to do when it's launched from your desktop?  As I understand it, the concept of MIME application/ types is that they emulate the situation where you open the file on your computer.  So on Windows, the application will look at the 3-letter extension; on the MAC, I assume there must be some way to look at the resource fork.  Many applications also parse the file for key header data (ie, an RTF file begins with {\rtf1, and so on).  Obviously, apps like Windows Media or MPlayer can play upwards of 50 different codec types, all with an .AVI extension (they look at the codec descriptor within the file).  Same thing with MoviePlayer for Quicktime.
> 
> Possibly a better approach would be the use of MIME parameters
> to indicate the codec, e.g.,
>   audio/ogg,code=vorbis

Again, I'm getting up to speed here, but my understanding is that these MIME types are for streaming media, NOT opening a file via HTTP.  In most cases, this means RTP streaming.  So the thinking appears to be, you better know which streaming client is going to parse your data before you start streaming it, since streaming is a real-time phenomenon (ie there is a timestamp, you may not have the ability to throttle the data, etc).

The present thinking appears to be: get the IETF to adopt application/ogg, with the idea that somewhere down the line there will be a media player that can handle all OGG types gracefully.  In addition, I believe we're asking for an audio/vorbis type as well, to cover audio streaming.  Presumably we will ask for a video streaming type as well at some point.

Suddenly I think I know why there is confusion wrt the use of the "Theora" label.  If we're going to get a MIME type to cover synchronized audio & video streaming using RTP, that type will need to have a name.  Theora-Vorbis is a bit of a mouthful.  Therefore there might be some arguable logic behind saying that we need a term for synchronized streams that specifically use Theora video and Vorbis audio.  OGG is inapplicable here since we've already defined it as a file wrapper and an application type, with a wide array of possible codecs within it.  It seems that Sylvia and Linus assumed the Theora label was the appropriate name for that stream.  ("Ogg Theora" multiplexes a Vorbis-encoded audio bitstream and a
VP3-encoded video bitstream in a single physical Ogg bitstream.").  This would lead to a MIME type of video/Theora.  If we don't like that naming terminology, we need to suggest something else, like video/oggvth (or whatever).
> 
> But the problem is that very few MIME processing applications
> (e.g., web browsers and file viewers) give you the kind of control
> to actually invoke the right thing.
> 
> Linus Walleij wrote:
> 
> > On Thu, 2 Jan 2003 Silvia.Pfeiffer at csiro.au wrote:
> > 
> > 
> >>Yet, I don't believe that application/ogg-tarkin or
> >>application/ogg-vorbis should be chosen, as "application/" 
> to me flags
> >>that the content of the file can only be determined by an 
> application
> >>that can read this specific type of file. In my opinion, 
> vorbis should
> >>be audio/ogg-vorbis because vorbis itself is an encoding format that
> >>makes sense in containers only, such as ogg or rtp.
> >>
> > 
> > I agree with Silvia. The above is also the view of the IETF. We want
> > audio/vorbis for Vorbis audio, no matter if it is inside an 
> Ogg or RTP
> > stream (or anything else for that matter, for example Quicktime)
> > 
> > audio/vorbis is to signify vorbis audio encoding no matter if it is
> > encapsulated in ogg or RTP or whatever IETF standard. Eg 
> audio/ogg-vorbis
> > would be highly inapropriate. This has been discussed with the
> > relevant IETF groups in the past, and also announced on the 
> IANA content
> > type mailing list.
> > 
> > Accordingly video/theora will most likely be used for 
> theora over Ogg
> > streams, RTP etc.
> > 
> > Any disagreements, please draft your own Internet Drafts 
> and add to the
> > IETF repository and we'll weigh them against each other and 
> have a chat
> > with considered parties.
> > 
> > Linus
> > 
> > 
> > 
> > 
> 
> 
> -- 
> 
> --- David A. Wheeler
>      dwheeler at ida.org
> 
> 
> --- >8 ----
> List archives:  http://www.xiph.org/archives/
> Ogg project homepage: http://www.xiph.org/ogg/
> To unsubscribe from this list, send a message to 
> 'theora-dev-request at xiph.org'
> containing only the word 'unsubscribe' in the body.  No 
> subject is needed.
> Unsubscribe messages sent to the list will be ignored/filtered.
> 
--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to 'theora-dev-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.