[
Lists Home |
Date Index |
Thread Index
]
-----Original Message-----
From: Dare Obasanjo [mailto:dareo@microsoft.com]
Sent: Sunday, July 28, 2002 6:57 AM
To: Clemens Vasters; xml-dev@lists.xml.org
Subject: [xml-dev] RE: On Schemas, Namespaces and Syntax vs. Semantics (long
but worth it) :)
> 2. XML Schema does not know versions, they know derivations.
>If you want to extend a Schema, create a new one, import or reference the
old one and restrict or extend the existing types.
Actually you can also use xs:redefine which allows one to extend the types
defined in a schema in a pervasive manner without resorting to changing the
targetNamespace of the schema.
Also note that substitution groups create yet another means of derivation;
based on extension/restriction, but which can affect structures not
explicitly extended.
On the topic of semantics. I must wade in, even though I fear that the
water's a bit too deep: we are indeed talking about data interchange and
about correctly identifying the version of data structure and type. Schema
can blur the line between syntax and semantics, simply due to its ability to
define and enforce more sophisticated type definitions. But we might be able
to agree that even richly defined and enforced structures and types are
still in the realm of syntax (that would have to be a matter of agreement,
since there is not absolute "truth" in these distinctions).
Given that agreement, there may still be the need to indicate, somewhere in
the interchanged data, the two parties' agreed-upon semantics for the
interchange. What I'm gleaning from this debate is that we can and should
distinguish between a schema (deep syntax) version and some other version
indicator that captures some snapshot of the versions of semantic
interpretation that the parties will employ. At very least, this will help
in situations where the schema (and thus the syntax) may be identical yet
the intended semantic interpretation different.
Your thoughts?
Mark
|
