Syncing up the MIME databases

Christian Rose menthos at menthos.com
Sun Nov 16 01:13:24 EET 2003


fre 2003-11-14 klockan 20.35 skrev Jonathan Blandford:
> > I've noticed some inconsistancies with these mime descriptions,
> > primarily with casing (some descriptions use lowercase, others mix
> > upper- and lowercase) and differences in the use of plural/singular.
> > 
> > Are there any plans to go through these descriptions and standardize
> > them with regards to things like this?
> 
> I didn't have any plans to do this, as I haven't really read through the
> strings.  I doubt there would be serious objections outside of the
> translators to doing this, though.

I could need some help with going through this though.

I found a document with the current gnome mime data description
guidelines in gnome-vfs/doc/mime-descriptions-guidelines.txt (also
attached). Could this also be of use for the common mime database?


Christian
-------------- next part --------------
gnome-vfs.keys file format:

	MIME types are listed in alphabetical order.

	although undesirable, it's unavoidable to sometimes use a duplicate
	description for separate MIME types.
	

general description format:

	if we know anything about the file, its english MIME description
	generally follows this form:

		[platform] [ [vendor] program name] <type> [($scheme-compressed)]
		e.g.: "Microsoft Word document", "tar archive (gzip-compressed)"

	if nothing is known beyond its compression, encryption, or encoding:

		[platform] <$scheme-$action> file
		e.g.: "Macintosh BinHex-encoded file", "PGP-encrypted file"

	extended description of components:

		platform:

			the platform name is included only if the file is
			intended to be used only on that system ("Macintosh
			BinHex-encoded file").

		program name:

			program names are included when the file format
			belongs specifically to that program, even if it can
			be read by other programs ("Microsoft Word
			document").

		type:

			the type is a lowercase word or short phrase
			("font", "source code"); we prefer the most specific
			accurate type.  when no appropriate type exists,
			"document" is a blanket term for files created by
			the user.  "file" is the least specific type.

			an "archive" is a file whose format may encapsulate
			other files.

			"image" and "video" are for formats suitable for
			photo-quality content; "art" and "animation" are
			those that are not.

			computer program code is called "<language> source
			code", unless there is another standard usage for a
			given language (e.g., "Perl script").


style issues:

	acronyms are capitalized and almost never expanded ("JPEG image").

	program and platform names are capitalized per conventional usage
	("FrameMaker", "gnumeric", "DOS").

	human language is preferable to acronyms;

	but short descriptions are necessary (so acronyms are often used
	anyway).

	the only whitespace allowable is single spaces between words.	


additional information:

	types which we used in the past and have been superceded include the
	following. the changes were made to near-synonyms to increase
	consistency.

		information	=>	data
		movie		=>	video
		graphic		=>	image

	"QuickTime movie" is a special case owing to common usage.


More information about the xdg mailing list