DOCBOOK: Re: DocBook filename extension

[Norman Walsh <ndw@nwalsh.com>]

/ "Matt G." <matt_g_@hotmail.com> was heard to say:
| meaningless.  Furthermore, is many cases, a file's filename is the
| only piece of information you have, about it.

I think that's probably the point at which to apply pressure. The
reason (or one of the reasons) that filenames are so important is
because that's (almost) the only metadata you can reliably extract
about a file if you don't know its format.

But once you've got a lot of your data in XML, you can imagine a tool
that extracts more metadata about a document than just its filename. I
can imagine being able to write Make-style rules that are based on
doctype or namespace name in addition to just filename.

Some of this could be shunted off into the filesystem. Why shouldn't a
filesystem be able to tell you the MIME type of a document, at least,
in addition to it's name and size and other properties? Watching
Windows try to associate applications with data files based on three
letter extensions should have tought us by now that there has to be a
better way. Extensions just aren't a big enough namespace (in the
non-XML sense) for the functionality we need.

I suppose if you imagine a fully-RDF'd (or topic-mapped, I'm agnostic)
world, you can imagine even fancier things.

My two, uh, cents.

                                        Be seeing you,
                                          norm

-- 
Norman Walsh <ndw@nwalsh.com>      | The man with ten children is
http://www.oasis-open.org/docbook/ | better off than the one with ten
Chair, DocBook Technical Committee | thousand fonts of type, because
                                   | the man with ten children doesn't
                                   | want any more.