Hello, I am working on expanding the capabilities of the MimeMagic module to reliably detect a larger number of media formats as part of my project to implement widespread support for video files. My experimentation on different file types as well as the existing amount of code dedicated to fixing the mistakes of facilities like `file -bi` have made me conclude that it would be worthwhile to modularize all the special-case code and supporting mime.type and mime.info data devoted to particular file types. I have nearly completed a mini framework to do that, which I'll be asking for a critique of soon. As I finish it up, though, I'd appreciate some clarification about how mapping of the various mime types to a media type (ie "AUDIO" or "ARCHIVE") was intended to work. Specifically, I can't figure out what the intent behind the "MULTIMEDIA" category is. By it's name it would seem to be a catch-all category, and some of the mime types assigned to it like application/ogg seem appropriate...but other members like video/quicktime and video/x-msvideo seem much more well defined. Was that an error, or is there some other deep dark purpose behind the "MULTIMEDIA" category that I should be aware of?
I ask because I plan to improve the accuracy of MimeMagic->getMediaType by consulting appropriate modules when necessary, and want to know if it is correct to do so ONLY if the type of the file in question fits into the MULTIMEDIA category.
Thank you, Mike
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Mike Baynton wrote:
Hello, I am working on expanding the capabilities of the MimeMagic module to reliably detect a larger number of media formats as part of my project to implement widespread support for video files. My experimentation on different file types as well as the existing amount of code dedicated to fixing the mistakes of facilities like `file -bi` have made me conclude that it would be worthwhile to modularize all the special-case code and supporting mime.type and mime.info data devoted to particular file types. I have nearly completed a mini framework to do that, which I'll be asking for a critique of soon.
Yay!
As I finish it up, though, I'd appreciate some clarification about how mapping of the various mime types to a media type (ie "AUDIO" or "ARCHIVE") was intended to work. Specifically, I can't figure out what the intent behind the "MULTIMEDIA" category is. By it's name it would seem to be a catch-all category, and some of the mime types assigned to it like application/ogg seem appropriate...but other members like video/quicktime and video/x-msvideo seem much more well defined. Was that an error, or is there some other deep dark purpose behind the "MULTIMEDIA" category that I should be aware of?
It looks like basically stuff-that's-not-quite-defineable-as-video.
It probably makes sense to reassign quicktime and AVI to the VIDEO category.
Ogg is slightly complicated because ogg files can contain audio, video, or both (or maybe something else entirely ;). Ideally we want a clearer way to distinguish between Oggs that are audio (just a Vorbis track) and video (with a Theora or possibly other video track). I'm not sure we have that in place yet.
Things like Flash probably could be reasonably classified as video... VRML... dunno. :)
- -- brion vibber (brion @ wikimedia.org)
wikitech-l@lists.wikimedia.org