> > Utilities like /bin/cp, /bin/tar, etc. will simply lose until someone
> > comes up with an API for accessing the metadata, so that those utilities
> > can deal with it.
>
> > Also consider other networking protocols besides NFS: AFS, FTP, HTTP,
> > etc. Implementing metadata really is a non-trivial exercise
>
> The Amiga Workbench used separate files with the extension .info in their
> names. That scheme is already compatible with everything you mention.
> Microsoft Windows uses a similar technique with its .pif (program
> information file) and .ico (icon) files, except that the Microsoft
> arrangement doesn't use them for hinting at which application should be
> used to operate on a data file--for that, the Windows shell keeps a list
> of applications, cross-referenced to the extensions usually given to the
> names of data files they generate.
This still doesn't handle multiple users all wanting different
"resources" for a single file. Say, you have /usr/bin/netscape
and one user wants the Mozilla icon and another user wants the
big N. This example is awful: you can think of better ones.
Resource forks need to be stored in your home directory. It is
not appropriate to look at single user machines, such as Amiga
and Macintosh and Windows, and then force their techniques for
resources into a multiple user environment like UNIX.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/faq.html