Namespace URI address resources

Rick Jelliffe ricko at
Thu Jun 17 16:50:57 BST 1999

From: David Megginson <david at>

>Murray Maloney writes:
> > Well. Maybe meta-data, but a schema is simply declarative.
> > It does not perform any processing. Editing, creating and
> > validating are all applications. So what?
>Schemas and stylesheets are both declarative, and both include
>information that can be acted on by processors.  In the case of a
>schema, the result is a truth value (valid/not valid) and, optionally,
>a transformation (supplying default values, etc.); in the case of a
>stylesheet, the result is a transformation or rendition.

I think that this exchange reveals the central confusion that people
have. If the editor of the Information Set draft and the editor of the
Schema draft still are feeling around the issue of a schema, what hope
for the rest of us :-) They are both exceptional men, but it shows that
there is lot of room for community discussion.

Murray  views schemas largely as the patterns by which documents are
constructed (or "described"). David views schemas as a specifications
for validation.

It is quite possible to build languages which do both, but the result
may be mediocre.  Xeena is OK but it is not great--even Adobe's
brilliant DTD-to-EDD transformations leaves a lot of things out that are
worthwhile for building.
In my view, DTDs are definitely validation schemas: they are a big fat
assert() statements bundled with other header information.

A schema that would be useful for document construction should include
human-readable comments, the ability of one attribute to enable another,
specifying the end types of reference,  etc.  Newbies ask predictable
questions about structures; the draft has not addressed those questions.

A language  validation should be some selection of the schema patterns
expressed as (possibly weaker) constraints and delivered in a very terse

There was a great book on parsing called "Syntax, Structures, Schemes,
Semantics, Verification".  It seems that people use "schema"
interchangably for any of these different things.

The XML Schema draft seems to be a "construction" schema for database
documents and simple untyped hypertext documents. It is certainly not a
validation suited for documents with structures outside tree structures,
and it is not suited for client-end validation unless there are no
bandwidth issues.

Rick Jelliffe

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at
Archived as: and on CD-ROM/ISBN 981-02-3594-1
To (un)subscribe, mailto:majordomo at the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at

More information about the Xml-dev mailing list