Talk:File URI I18N

Terminology

Web of material

"A web of material" and the notion of "moving a web" to a local file system are phrases I haven't heard before. What did you mean exactly? Are you talking about mirroring a collection of resources? I'd prefer to use those terms. - Mjb 12:58, 27 Dec 2006 (MST)

File name, file path, OS path

You had "file name" but I changed it to "OS path". People have different ideas about what these terms mean. When you say "file", you seem to exclude folders/directories, which could well be the resources being referenced (e.g., C:\Program Files is a valid OS path that can be mapped to a file: URI, but it probably refers to a folder, not a file). It also implies that you're talking about just the trailing component of the path, whereas for purposes of mapping to and from URIs we need to make sure people think of the whole path as a unit. Further complicating things is the fact that URI has a path component, so we want to avoid using the word path too ambiguously. So I suggest we settle on some definitions. How do you feel about this, below?

An OS path is an operating system and/or file system-specific designation for a single resource, as a sequence of characters. A file URI identifies a resource by representing an OS path. If the OS path does not unambiguously identify a single resource, it cannot be mapped to a file URI. However, the same resource may be identified by multiple OS paths.

I think we should stop there, but we could go on to clarify that an OS path conceptually encompasses all of the other components one might be tempted to consider to be irrelevant. I'd say something like this: Generally, (this is not prescriptive and doesn't bear on how the path is mapped to a file URI) an OS path consists of at least two components: a file name, which is a character-based name, as opposed to an inode ID; and a path, which is a character-based string providing contextual information for the purpose of differentiating the resource from others with the same name, typically by describing a chain of "directories" or "folders" in a hierarchical organization system. A directory/folder may be the resource identified by an OS path; in this case, its name is the "file name". The path component may have subcomponents - for example, MS-DOS/Windows paths may consist of a directory path preceded either by an optional drivespec, consisting of a drive letter followed by colon ("C:"), or, in the case of UNC paths, by an optional computer name and share name ("\\fileserver\backups"). Other file systems may use terms like like "volume" and "record".

The following diagram of a Windows file system path demonstrates that the OS path conceptually encompasses all of the other components:

 C:\Program Files\Internet Explorer\iexplore.exe
 \/|\___________/|\_______________/|\______/ \_/
 | |     |       |        |        |    |     |
 4 5     6       5        6        5    7     8
 \_________________________________/\__________/
                 |                        |
                 2                        3
 \_____________________________________________/
                       |
                       1

1. OS path - identifier of a resource in a file system - this is what gets mapped to a URI
2. path - a file system specific component of the OS path
3. file name - a file system specific component of the OS path
4. drivespec - a file system specific subcomponent of the OS path
5. separator - a file system specific subcomponent of the OS path
6. directory name - a file system specific subcomponent of the OS path
7. base name - a file system specific subcomponent of the OS path
8. extension - a file system specific subcomponent of the OS path

- Mjb 12:58, 27 Dec 2006 (MST)