RE: Default and Fixed Values of attributes

Dear Michael,
 
We get to the point: what IS a default value?
 
On Google define:  I found: 
1.	"selected automatically unless an alternative is specified"
2.	"when a physically meaningful value cannot be computed a default
value is provided"
3.	"A value a field will assume unless an explicit value is entered
for that field"
4.	"A predetermined, frequently used value for a data field or
control entry, intended to reduce required user entry actions"
5.	"A value that is automatically displayed by the computer. Users
can often override the default by deleting the old value and entering a
new one. The goal is to choose a value that will almost always be
entered, so the user can skip that item"
and so on.
 
That is also what I understand it to be. So when the base type attribute
is "required" and the inherited subtype attribute gets a default value,
that default value will be used when nothing would be filled in, and the
"required" constraint would be fulfilled. So is it possible that this is
a flaw in XML Schema (if I dare say such a blasphemous thing) ? Because
if it isn't, one MUST make the use="optional" for attributes of any
complexType that MAY, somewhere, have a subtype with a default value,
and that is unmanageable and unforseeable. We'll have the schemas
accessible on the Internet, for import in user-defined schemas.
 
Regards,
Hans
 
========================================================================
==========================
-----Original Message-----
From: xmlschema-dev-request@w3.org [mailto:xmlschema-dev-request@w3.org]
On Behalf Of Michael Kay
Sent: Monday, February 14, 2005 3:19 PM
To: 'Hans Teijgeler'; 'XML-Schema Developers Forum'
Cc: 'Paap, Onno'
Subject: RE: Default and Fixed Values of attributes


If the base type makes the attribute "required", that means the
attribute must be present in every instance of the type. You can't make
it optional in a subtype, because if you did, it would allow instances
of the derived type to be invalid against the base type. And if it's not
optional, there's no sense in having a default value.
 
Michael Kay
http://www.saxonica.com/

  _____  

From: Hans Teijgeler [mailto:hans.teijgeler@quicknet.nl] 
Sent: 14 February 2005 14:04
To: XML-Schema Developers Forum
Cc: Michael Kay; Paap, Onno
Subject: FW: Default and Fixed Values of attributes


Dear Michael,
 
I downloaded Stylus Studio's XML Editor and validated the same file, and
also had the same Catch 22 when a Default value is entered:
*	the use MUST be consistent with the base type 

*	since the base type attribute has use="required" I changed that
back to that, BUT: 

*	then I get the error message: "If default and use are both
present, use must have the value 'optional' "
 
Can you or someone else help me to get out of this booby trap?
 
Regards,
Hans
 
-----Original Message-----
From: xmlschema-dev-request@w3.org [mailto:xmlschema-dev-request@w3.org]
On Behalf Of Hans Teijgeler
Sent: Monday, February 14, 2005 12:38 PM
To: 'Michael Kay'; 'XML-Schema Developers Forum'
Cc: 'Paap, Onno'
Subject: RE: Default and Fixed Values of attributes


Dear Michael,
 
Thanks for your blazingly fast response, written on a Sundaynight! It
gives some clues for further digging.
 
1)   In 2.1 it says that the stuff in 2.1.3 must be true IF B is in the
same target namespace as R. That is interesting. What if the schema that
defines R is imported in the schema that defines B?
 
2)   I would be obliged to receive your reaction on the other questions
in my message, viz:
*         Whether or not the definition of a default or fixed value
should have as a consequence that the use constraint should become
"optional" (for default values) or even should be removed at all (for
fixed values). 
*         Can you explain the rationale for the split between
specialization(typing)_by_extension and specialization_by_restriction?
>From a data modelling point of view that is illogical, because adding an
attribute or child element also constitutes an extra constraint, and
besides that it is highly inconvenient and causes a lot of hassle.
 
Kind regards,
Hans
 
========================================================================
===============================================
 
-----Original Message-----
From: xmlschema-dev-request@w3.org [mailto:xmlschema-dev-request@w3.org]
On Behalf Of Michael Kay
Sent: Sunday, February 13, 2005 6:50 PM
To: 'Hans Teijgeler'; 'XML-Schema Developers Forum'
Cc: 'Paap, Onno'
Subject: RE: Default and Fixed Values of attributes
 

  _____  

From: xmlschema-dev-request@w3.org [mailto:xmlschema-dev-request@w3.org]
On Behalf Of Hans Teijgeler
Sent: 13 February 2005 16:38
To: XML-Schema Developers Forum
Cc: Paap, Onno
Subject: Default and Fixed Values of attributes
Ladies and gentlemen,
 
I am struggling with Default and Fixed Values for attributes, and I need
your advice.
 
Assume:
1. complexType A with attribute 'a' 
2. complexType B is typed with A and the inherited attribute 'a' is
restricted by means of a default or fixed value "value1" 
3. complexType C is typed with B and the inherited attribute 'a' has the
inherited default or fixed value "value1" 
 
Now I want to define an other default or fixed value "value2" for
inherited attribute 'a' of complexType C because that value is, in the
real world, different for this subtype. WRONG!!! When I do that (in
XMLSpy, latest version) that "value2" is kind of "inherited backwards"
to the inherited attribute 'a' of complexType B, but not to A (see
below) 
 
The rules for "fixed" and "default" are slighly different. For "fixed",
Schema Part 1 secton 3.4.6,  Schema Component Constraint: Derivation
Valid (Restriction, Complex) tells you:
 
 2 For each attribute use (call this R) in the
<http://www.w3.org/TR/xmlschema-1/#ct-attribute_declarations> {attribute
uses} the appropriate case among the following must be true:
2.1 If there is an attribute use in the {attribute
<http://www.w3.org/TR/xmlschema-1/#ct-attribute_declarations> uses} of
the {base  <http://www.w3.org/TR/xmlschema-1/#ct-base_type_definition>
type definition} (call this B) whose {attribute
<http://www.w3.org/TR/xmlschema-1/#attribute> declaration} has the same
{name} <http://www.w3.org/TR/xmlschema-1/#a-name>  and {target
<http://www.w3.org/TR/xmlschema-1/#a-target_namespace> namespace}, then
all of the following must be true: 
2.1.3 [Definition:]  Let the effective value constraint of an attribute
use be its {value
<http://www.w3.org/TR/xmlschema-1/#au-value_constraint> constraint}, if
present, otherwise its {attribute
<http://www.w3.org/TR/xmlschema-1/#attribute> declaration}'s {value
<http://www.w3.org/TR/xmlschema-1/#a-value_constraint> constraint} .
Then one of the following must be true: 
2.1.3.1 B's  <http://www.w3.org/TR/xmlschema-1/#key-evc> .effective
value constraint. is  <http://www.w3.org/TR/xmlschema-1/#key-null>
.absent. or default.
2.1.3.2 R's  <http://www.w3.org/TR/xmlschema-1/#key-evc> .effective
value constraint. is fixed with the same string as B's. 
 
In other words, if B has a fixed value then R (your C) must have the
same fixed value. That's reasonable: a restricted subtype can only
permit values that are valid for the supertype.
 
On the other hand, if B has a default value there is no rule that R
(your C) must have the same default value.
 
However, there's nothing in the spec corresponding to the "backwards
inheritance" you describe. 
 
 
 
XMLSpy seems to be internally confused 
 
I think this is the nub of your problem. XML Spy is not the most
conformant of schema processors. If you want to use trial and error to
find out what the spec means (and we all need to, occasionally) then use
Xerces and XSV: if they give the same answer, it's almost certainly
right.
 
Michael Kay
http://www.saxonica.com/