>As Daniel points out, XML, with its insistence on systems identifiers, is
>not particularly friendly to the notion of the SGML open catalog in
>particular or in general to the the concept of identifying external
>resources as general entities.

This is not the first time that I've seen XML's "insistence on SYSTEM
identifiers" referred to. The XML standard allows PUBLIC identifiers to be
included, and does not force them to be ignored. It does insist that
processors that _can't_ understand PUBLIC IDs should use the supplied
system ID. Some of fought hard against this, but you can't win them all.

There is nothing to prevent the use of PUBLIC identifiers and catalogs
except the unwillingness of users to demand support from vendors.

Of course, delivering XML straight to Explorer is not going to work without
working system IDs, but there's no reason not to use XML tools for archival
as well as delviery, as long as one of the production steps is the
modification of SYSTEM IDs to something sensible.

In fact, in the long run, it seems to me that XML is a better archival
format since it's so much easier to parse and process. This makes it more
robust, because it's not hard to create a tool even if a good one isn't
available. I think that in the long run the tool support is going  to
improve faster because there will be more competing vendors.

  -- David
David Durand              [log in to unmask]  \  [log in to unmask]
Boston University Computer Science        \  Sr. Analyst   \  Dynamic Diagrams
MAPA: mapping for the WWW                    \__________________________