W3C home > Mailing lists > Public > www-xml-schema-comments@w3.org > July to September 2007

[Bug 3221] Terminology: "value"

From: <bugzilla@wiggum.w3.org>
Date: Mon, 17 Sep 2007 20:28:40 +0000
CC:
To: www-xml-schema-comments@w3.org
Message-Id: <E1IXNCy-0002zM-1m@wiggum.w3.org>

http://www.w3.org/Bugs/Public/show_bug.cgi?id=3221





------- Comment #2 from cmsmcq@w3.org  2007-09-17 20:28 -------
Owing to travel, I was absent when this issue on the usage of the 
term "value" was discussed.  

Had I been present on the call, I think I would have argued that the premise 
of the comment may need discussion or clarification.  The text of the spec 
does not make it very explicit (perhaps it should do so), but over time, the
XML Schema WG has come to agree that the best way to interpret the XSDL
spec is to assume that given any value, one necessarily is also given the
identity of its primitive datatype.  So that one knows, given a sequence of
bits, whether it is a value of type xs:hexBinary or of xs:base64Binary. 
This seems in many ways analogous to the idea in the QT specs that when 
given a value, one knows what type it has.  The difference is that XSDL
assumes one knows the primitive datatype, while in QT one knows the
governing type definition (or the active member type definition.  

But it is true that XSDL regards the identity of the primitive type
as just a property of the value, rather than talking about pairs
whose one member is the name of a type and whose other is the 'value' 
qua datapoint within the value space of that type.

It would probably be worth while, both in XSDL and in the relevant
QT specs, to point to this difference in usage and to point out that
the difference in viewpoint is not, in practice, all that great.
Received on Monday, 17 September 2007 20:28:42 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:50:06 UTC