Re: Simple Type definition Language

At 10:07 AM +0000 2007-04-12, you wrote:
>Ladies amd Gentlemen,
>
>In part 2 chapter 3 of XML Schema,  "Built-in datatypes", a 
>mechanism is described to reference a Schema datatype via am URI.
>For example a simple int is referenced via 
><http://www.w3.org/2001/XMLSchema#int>http://www.w3.org/2001/XMLSchema#int.
>Facets may be referenced, too, e.g.: 
><http://www.w3.org/2001/XMLSchema#maxInclusive>http://www.w3.org/2001/XMLSchema#maxInclusive
>
>But nothing is said how this mechanism could be used to defined a 
>restricted data type.

>Here is my proposal:
>
>An restriction definition is attached to the URI within squared 
>brackets in a XPath-like style, e.g.:
><http://www.w3.org/2001/XMLSchema#int[minInclusive=0>http://www.w3.org/2001/XMLSchema#int[minInclusive=0 
>and maxInclusive=100]
>Now the value is restricted to the interval 0..100.
>
>Round brackets could be used for more complex restrictions:
><http://www.w3.org/2001/XMLSchema#int[(minInclusive=1>http://www.w3.org/2001/XMLSchema#int[(minInclusive=1 
>and maxExclusive=5) or (minInclusive=10 and maxExclusive=15)]
>Now the value may be 1<=x<5 or 10<=x<15.
>
>To define string enumerations one could use commas:
><http://www.w3.org/2001/XMLSchema#string[enumeration='val1','val2','val3'>http://www.w3.org/2001/XMLSchema#string[enumeration='val1','val2','val3']

The schema working group considered your suggestion on 17 August.

We accept the goal, and think the proposed mechanism is ingenious, but
isn't complete (unions, complex types) and doesn't offer enough benefits
over SCDs to be proposed right now.

The WG thinks that this is an ingenious solution, but declines to make the
change to this version of datatypes.

This matter is tracked at the W3C bugzilla system, as bug 4886:
   http://www.w3.org/Bugs/Public/show_bug.cgi?id=4886
If you accept this decision of the WG, we would appreciate if you would
mark that issue closed; alternatively, if you wish to object, please
mark it reopened.  If you are unfamiliar with bugzilla, you may send
a message to the comments list (where you sent the original suggestion)
stating your desire and we will mark it on your behalf.  In any case,
if we do not receive a reply within three weeks, we will assume that
you accept the decision and will close the bug.

We do appreciate your effort, and it may yet develop into something
in a subsequent release of the spec.  Accordingly, we have marked the
bug as "later" rather than "won't fix".

Thank you for caring about the Schema specification.
-- 
Dave Peterson
SGMLWorks!

davep@iit.edu