XCatalog proposal draft 0.1

[John Cowan]

Murray Altheim scripsit:


> You speak of an XML world as if it existed as a fixed entity.

No, but one part of it is fixed: the spec.

> [I]f people wish to build
> large systems based on XML, I'm suggesting we look at the reasons why 9401
> has its feature set and use this to inform our design.

As far as I can see, the reason is that SGML doesn't give any
definite meaning to SystemIds, and some kind of outside support
is needed to map them to filenames or other things that
programs understand.  XML *does* define the meaning of SystemIds,
by reference to various URI documentation.

> There very well
> may be a need for URI-to-URI mapping in some systems. I'm simply
> advocating that before we attempt to 'standardize' XML catalog
> resolution, we're going to need a lot more input, hence my suggestion
> of an IETF Internet Draft.

Right.  Since both the Socat and the instance format are highly
extensible (as explained in my earlier posting), there is no
problem with standardizing only some features and adding others later.
 
> As I stated in my earlier message, my posting is not meant to be critical
> (other than bring up technical points) of the XCatalog proposal. I think
> XML desperately needs catalog (ie., name resolution) support, both because
> I am against hardwiring locations in documents,

And I am afraid of preempting URN resolution mechanisms, because
if you don't like hardwired URLs, then URNs are probably what you
want.

In utter truth, PublicIds are just another kind of URN,
but they are already defined and URNs are not.  I want to start
creating a PublicId infrastructure analogous to the DNS
infrastructure, but piggybacked on existing HTTP servers.
A model implementation of a PublicId to SystemId resolver
would go a long way to making such a thing easy for people to
provide.  Publish a few XML documents, add a few lines to the
catalog.

To my mind, local remapping facilities are secondary to that goal.
In fact, in a regime of public catalogs, they would be downright
undesirable: consider the effects of a catalog that
remaps somebody else's URLs to your own site.  That would subvert
the current (moderately strong) assurance that "http://www.sun.com/..."
*semper ubique et ab omnibus* refers to a page on Sun's web site.

> and because without it,
> XML may end up being simply an output format for existing SGML systems.

I don't understand this point.  There are surely already many XML
formats that are by no means derived from SGML system output.

> You seem critical of the work of standards bodies as being too slow.

"Slow" and "too slow" are surely not synonymous.

> I think that this might also be proposed as a work item for OASIS, and
> I'll poke around to see if this might already be in anyone's neural fibre.

Thank you.

-- 
John Cowan	http://www.ccil.org/~cowan		cowan at ccil.org
	You tollerday donsk?  N.  You tolkatiff scowegian?  Nn.
	You spigotty anglease?  Nnn.  You phonio saxo?  Nnnn.
		Clear all so!  'Tis a Jute.... (Finnegans Wake 16.5)

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)