W3C home > Mailing lists > Public > xmlschema-dev@w3.org > October 2012

Re: "Unsafe" union?

From: Michael Kay <mike@saxonica.com>
Date: Tue, 16 Oct 2012 15:15:17 +0100
Message-ID: <507D6BF5.5010507@saxonica.com>
To: xmlschema-dev@w3.org
The principle of type safety is the rule that if X is substitutable for 
Y (i.e. an instance of type X can be supplied where an instance of type 
Y is expected), then every valid instance of X must be a valid instance 
of Y.

if you define a union type U as union of xs:date and xs:dateTime, then 
xs:date and xs:dateTime are substitutable for U: for example an element 
of type xs:date can safely be a member of the substitution group of an 
element of type U because every instance of xs:date is a valid instance 
of U.

However, if you define another union type V as a restriction of U, with 
the constaining facet <pattern value="20.*]"/>, then not every xs:date 
is a valid instance of V (specifically, dates outwith the 21st century 
are not). However, XSD 1.0 incorrectly treats xs:date as being 
substitutable for V, which means that a program expecting to process 
21st century dates might find itself with a date from a different 
century on its hands. XSD 1.1 fixes this error, and XPath 3.0 
incorporates this fix.

Michael Kay
Saxonica


On 16/10/2012 12:08, Costello, Roger L. wrote:
>
> Hi Folks,
>
> The XPath 3.0 specification has this statement:
>
>     The current (second) edition of
>
>     XML Schema 1.0 contains an error
>
>     in respect of the substitutability of
>
>     a union type by one of its members:
>
>     it fails to recognize that this is unsafe
>
>     if the union is derived by restriction
>
>    from another union.
>
> Huh?
>
> What is "unsafe"?
>
> What is that saying please?
>
> The XPath 3.0 specification then says:
>
>     This problem is fixed in XSD 1.1, but
>
>     the effect of the resolution is that an
>
>     atomic value labeled with an atomic type
>
>     cannot be treated as being substitutable
>
>     for a union type without explicit validation.
>
> Again, what is that saying please?
>
> /Roger
>
Received on Tuesday, 16 October 2012 14:15:46 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 23:16:02 UTC