Re: [ubl] For review: Revised Version *Element* Rules

["G. Ken Holman" <gkholman@CraneSoftwrights.com>]

At 2005-11-09 12:38 -0500, Grimley Michael J NPRI wrote:
>     Every UBL document schema MUST include a required element named 
> "UBLVersion" as the first child of it's root element. This element 
> MUST have a default value that matches the value of the xsd:version 
> attribute of its containing schema.

I see the version of the schema as a property or characteristic of 
the instance and not a data element of the document element.  I know 
there is quite the religion around whether an information item should 
be an attribute or should be an element, but for something like a 
version characteristic, I feel strongly this should be an attribute.

And I think that default values in schemas should be avoided and the 
instance rule is sufficient:

>Instance Rule:
>
>     The value of the 'UBLVersion' element MUST match the value of 
> the xsd:version attribute of the UBL Document schema it instantiates.

Right (except I would have said the UBLversion= attribute) ... which 
means it is irrelevant what the schema declares the attribute to be 
as a default value, it is only relevant that the attribute is allowed 
to exist.  Indeed it is redundant to require the schema to default a 
value since the instance is obliged to have a value.

In my ISO work with DSDL I've come to believe that the role of 
grammar and type schemas is to constrain structure and the role of 
assertion schemas to constrain values.  A default value is not a 
property of the structure, rather, it is a particular application's 
use of that structure and different values might apply to different 
applications.

I've come to believe an instance should be allowed to live entirely 
on its own, well-formed, without a schema and without a schema 
location, and if we were to need to introduce any default values 
implied by the schema then instances would never be allowed to be 
standalone.  Processing should require the schema in order to work 
(as is true for XSLT 1.0 and rather not true for much of the use of XSLT 2.0).

I hope this helps.

. . . . . . . . Ken

--
Upcoming XSLT/XSL-FO hands-on courses:  Denver,CO March 13-17,2006
World-wide on-site corporate, govt. & user group XML/XSL training.
G. Ken Holman                 mailto:gkholman@CraneSoftwrights.com
Crane Softwrights Ltd.          http://www.CraneSoftwrights.com/o/
Box 266, Kars, Ontario CANADA K0A-2E0    +1(613)489-0999 (F:-0995)
Male Cancer Awareness Aug'05  http://www.CraneSoftwrights.com/o/bc
Legal business disclaimers:  http://www.CraneSoftwrights.com/legal