Wikipedia talk:Naming conventions (file formats)

Let's work out a convention for articles on computer file formats.

I'm not happy about "TAR file format". Firstly, many people advocate writing file extensions in lowercase. Secondly, it's a pain to use, since most computing articles will only want to write "tar".

I suggest one of:
 * "tar (file format)"
 * "tar (computing)"
 * "tar (format)"

(this is of course only for those names that require disambiguation. Pages like "PNG" and "JPEG" don't need anything in parentheses)

-- tarquin

(by the way, I've just noticed on Talk:Tar that it was ME who suggested the "TAR file format" article name. Aaaaaarg. Have I been sleepwalking or something? -- Tarquin 11:44 Jan 11, 2003 (UTC))

"tar (file format)" is probably best. Upper vs lower case is a tricky one: some extensions are acronyms which implies upper case, but on the whole, lower case seems better. I think it's more common at any rate. Tannin

Proposal 1
by Quuxplusone (talk)

I propose the following guidelines:

Guidelines

 * 1) There should be a family of pages titled .foo (file extension), each of which should be listed under Category:File extensions. These pages should be
 * 2) Redirects, in the case of unambiguity (.png (file extension)&rarr;PNG, .bmp (file extension)&rarr;Windows bitmap).
 * 3) Disambiguation pages, in the case of ambiguity (.doc (file extension), .cas (file extension)).
 * 4) Main articles will fall into at least three categories:
 * 5) Articles about file formats independent of specific programs (PNG, MOD (file format), ZIP (file format)). These should be named as per those examples: all-caps, with a "(file format)" dab if needed.
 * 6) Articles about programs with dedicated formats or extensions (gzip, tar (file format)). These should be named according to common usage, which will vary between *nix and non-*nix platforms.
 * 7) * *nix: see Category:Unix software. The general form is foo, disambiguated as needed with foo (Unix) or foo (mimsy frobulator) for programs that have achieved wide notability outside the *nix community.
 * 8) * Software targeted at non-*nix cultures usually has a brand or trade name: Adventure Game Toolkit, Microsoft Word. Use it.
 * 9) Articles about common usage of extensions (Readme, .txt)
 * 10) Only main articles should be tagged with categories such as Category:Archive formats. Redirects should never be placed in categories, with the special exception of Category:File extensions as described above.

Rationale

 * 1) The "File extensions" category can then serve as a Wikipedia GFDL mirror of dictionaries like wotsit.org. Redirects from e.g. .png may also be created, but not listed in that category, so there are no duplicates. The reason for the "(file extension)" dab is that otherwise we'll have to make exceptions for .com, .us, and so on. I think it's best to have everything symmetric.
 * 2) (As of this writing, gzip has been moved to GNU zip for reasons unknown and un-discussed. I'm hoping it moves back soon.)
 * 3) * By this logic, tar (file format) should be at tar (Unix). But no moves should happen based on this proposal until and unless it is accepted by a fair number of editors.
 * 4) * Use the Frobozz Wuffalo (wuffulator) style of disambiguation if needed &mdash; but it really shouldn't be needed. Version differences can be handled ad hoc; for example, Frobozz Wuffalo 3 or Wuffalo (version 9).
 * 5) For example, ".txt is a file extension usually associated with plain ASCII text." But there is no file format associated with that extension. Contrariwise, one file format using ASCII text is .csv (file extension)&rarr;Comma-separated values.
 * 1) * Use the Frobozz Wuffalo (wuffulator) style of disambiguation if needed &mdash; but it really shouldn't be needed. Version differences can be handled ad hoc; for example, Frobozz Wuffalo 3 or Wuffalo (version 9).
 * 2) For example, ".txt is a file extension usually associated with plain ASCII text." But there is no file format associated with that extension. Contrariwise, one file format using ASCII text is .csv (file extension)&rarr;Comma-separated values.

Counterproposal

 * There is rarely anything to say about file extensions; instead, these should in almost all cases be redirects to an article about the file format, a list of file formats (e.g. for the many text-like formats) or a program using the file format (the sole or most common program, e.g. ".doc" &rarr; Microsoft Word).
 * Article names should not start with a dot.
 * Redirect pages should not be in categories.
 * The extension "tar (file format)" should only be used when it's necessary because of the term (tar) having other meanings. "(file format)" is preferable over naming an operating system using it, unless there are different operating systems that use it in a different way.
 * There is rarely a need for separate articles on new versions of a program. Merging may be appropriate.
 * Radiant_ &gt;|&lt; 12:09, August 25, 2005 (UTC)