RE: PSVI

["HUGHES,MARK (Non-HP-FtCollins,ex1)" <mark_hughes@non.hp.com>]

>From: Tim Bray [mailto:tbray@textuality.com]
>Another example is from a post here by Andrew Layman from 16 
>September 1999:
>  Elements defined by a schema, when used in an instance document 
>  in a validating processor, will have these default values available, 
>  and this fact is pertinent to the author of the document.  This 
>  means that an element is incompletely read if the schema for it is 
>  not read. 
>This statement is at the very least controversial.  Are there a 
>others around here who would defend this point of view?   I 
>apologize if I've quoted Andrew out of context; the words above
>may not represent his feelings, but it is a good example of the
>schema-centric view of reality.  -Tim

  Right, Andrew's position is very clearly not the case in "real-world XML"
(at least for my value of "real-world").  You can gain additional,
secondary,
information by processing a schema for an XML document, but the document
contains perfectly usable and complete information if all you want are
nested
elements (often without attributes), and text nodes, and you can trust your
software to produce valid documents (this is what software testing is
for...)

  We do need to have clear specifications and divisions between "minimal
XML"
(merely well-formed documents, not even concerned with namespaces),
"namespace-qualified XML", "validating XML" (DTDs, schemas, etc.), and
higher
levels.  All of those are potentially valid uses, depending on your
application's needs.

  XML-RPC is a good example of minimal XML: there is no need for namespaces
or
schemas, because applications are sending it to well-known counterparts, in
a
well-known format (though it would be nice if Dave would write a more
precise
RFC-style spec for it instead of the conversational one), and there are no
extensions to need separate identification.  If this is not "complete", I'll
eat my hat.

  DocBook also gets by with just DTDs, and has just barely started to
stretch
into schemas.  It doesn't really need either, except to precisely specify
the
valid element structures.

  SOAP does need namespaces and schemas, for the data types as well as
including element structures from various sources, but you don't actually
have to validate a SOAP document to do useful things with it.

-- <a href="http://kuoi.asui.uidaho.edu/~kamikaze/"> Mark Hughes </a>