> Sylvia Pfeiffer said: > > > Now, leading on from here to file extensions is a different issue. I > > would recommend to have different file extensions for each of the > > defined Ogg media mappings ogg-theora, ogg-vorbis and ogg-speex. My > > reasoning is that it makes it easier for programs to map > from the file > > extension to the MIME type and from there to the application to use > > without having to look inside the file. My check of MIME types also > > found that mostly a MIME type maps to many file extensions, > but one file > > extension belongs mostly to one MIME type only.I don't see how this helps, since the MIME type by definition calls one and only one application. Either that application can support the standard OGG codecs (Vorbis, Speex, Flac, Theora...), or it can't. If it can, it doesn't need a plethora of file extensions to do so. That information is (or certainly should be!!) encapsulated into the OGG header as some sort of descriptor, ala Quicktime & AVI ("VORB", "SPEX", "FLAC", "THEO"). The point of a format like OGG is to encapsulate these issues into the format; wrangling that info out to the file extension seems just plain wrong, especially if you're not going to have a different MIME type for each (ie a Vorbis player, a Speex player...) I would make exactly one exception to this rule: we should define .ogv to m --- >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 'vorbis-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
jeez, Outlook sucks. AS I was saying:> I would make exactly one exception to this rule: we should > define .ogv for any OGG file with a video stream. It should also be a separate MIME type, application/ogv. The thinking is this: it is reasonable for an application to be focused on playing audio only or video and audio synchronized. It is unreasonable to expect an audio player to play video. Therefore it would make sense to have a different MIME type for these two types of applications, and a different extension (for the same reason when the files sit on your hard drive).I think in general it seems to make the most sense to have a one-to-one mapping of extensions to MIME types (for application/ types at least). They pretty much serve the same purpose: to decide which application is capable of handling the file or http stream. An http stream is just a file on the server. It's always possible for a user to map multiple MIME types to the same application; but if there is only one MIME type, they have to set it to just one application. -dbm --- >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@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
>> Sylvia Pfeiffer said: >> >> > Now, leading on from here to file extensions is a different issue. I >> > would recommend to have different file extensions for each of the >> > defined Ogg media mappings ogg-theora, ogg-vorbis and ogg-speex. My >> > reasoning is that it makes it easier for programs to map >> from the file >> > extension to the MIME type and from there to the application to use >> > without having to look inside the file. My check of MIME types also >> > found that mostly a MIME type maps to many file extensions, >> but one file >> > extension belongs mostly to one MIME type only. > >I don't see how this helps, since the MIME type by definition >calls one and only one application. Either that application can >support the standard OGG codecs (Vorbis, Speex, Flac, Theora...), >or it can't. If it can, it doesn't need a plethora of file >extensions to do so. That information is (or certainly should >be!!) encapsulated into the OGG header as some sort of descriptor, >ala Quicktime & AVI ("VORB", "SPEX", "FLAC", "THEO"). The point >of a format like OGG is to encapsulate these issues into the >format; wrangling that info out to the file extension seems just >plain wrong, especially if you're not going to have a different >MIME type for each (ie a Vorbis player, a Speex player...)I support this point of view. Encoding meta-information in an arbitrary label external to the actual data, i.e. extension, is a bad idea from any point of view. One of the big mistakes of MS-DOS was this idea, even though better ideas were already in use (i.e. magic numbers, which although imperfect is so much better). We're still suffering from that decision more than two decades ago. If it's at all possible to avoid propagating this error into the sub- domain of Ogg, please spend some effort to achieve this! Sometime in the distant future, we may actually see all major operating systems and environments supporting the migration away from this legacy errant thinking. If a new format specification has the ability to ease that transition it's a good thing. File extensions are evil. IMHO. /Svante --- >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 'vorbis-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.