RE: What would break if ... ?

Which of the four xdt-types do you mean?

 

xdt:yearMonthDuration/xdt:dayTimeDuration: xsd:string is not a subtype
of xdt:duration. So that would not work.

 

xdt:untypedAtomic: This is the type used to denote dynamically instances
that have no more specific type than xs:anySimpleType or untyped values
(mainly from attributes). It then describes the static type of these
instances and is used in type expressions to refer to untyped atomic
values. Making this to be the same as xs:string would mean that we 

 

1.	either lose the ability to perform weak-typing on untyped data
since xs:string is strongly typed and cannot be used when the expected
type does not allow xs:string, 
2.	or make xs:string weakly-typed so that we continue to have the
implicit casts needed to operate on untyped data.

 

The working group considered these two alternative options and decided
that neither is acceptable for all people.

 

xdt:anyAtomicType: This is just an abstract type that fills a hole in
the type hierarchy. We used to have a keyword for that, but decided that
having a type name would be better.

 

I hope this answer is satisfactory.

 

Best regards

Michael

 

-----Original Message-----
From: Svgdeveloper@aol.com [mailto:Svgdeveloper@aol.com] 
Sent: Tuesday, May 13, 2003 2:59 AM
To: public-qt-comments@w3.org
Subject: What would break if ... ?

 

I guess that at least several of the members of the WGs are aware of
discussions ongoing on XML-Dev and on XSL-List.

One question which I would like to pose to the WGs is, "What would break
if we abolished the xdt types and just allowed xsd:string in those
places where xdt types are currently proposed?".

I guess I could embellish the question with lots of further questions
and comment but it would help me, and possibly others, to understand
what might happen with that possible change.

Andrew Watt