Wildcards in SML schema

[Grant Robertson]

If I use the 'any' element in my schema to allow elements from another
schema to be used in instance documents based on my schema, is there a
way to force that the contents of that element must be an entire,
complete instance document for that other schema?

Let's say I have the following schema:

<?xml version="1.0" encoding="UTF-8"?>
<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema">
<xs:element name="BogusElement" type="BogusType"/>
<xs:complexType name="BogusType">
<xs:sequence>
<xs:any namespace="http://www.w3.org/1999/xhtml"/>
</xs:sequence>
</xs:complexType>
</xs:schema>

It has only the one element BogusElement and it can contain only the one
child element from the XHTML schema. Does this force the BogusElement
element in the instance document to contain a complete XHTML document
including headers or would

<?xml version="1.0" encoding="UTF-8"?>
<BogusElement
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:noNamespaceSchemaLocation="Untitled1.xsd"
xmlns:xhtml="http://www.w3.org/1999/xhtml">
<xhtml:strong>Bolded Text.</xhtml:strong>
</BogusElement>

be a valid document under the above schema?

How can I make it so that the contents of the BogusElement element in the
instance document must be a complete XHTML document in order to be valid?
Or would I just have to make that a verbal rule, not expressed in the
schema?

To complicate things further: Let's say I wanted that element to contain
only one of a list of valid, complete documents such as either a complete
html document OR a complete XHTML document OR a complete DocBook
document, all with headers. How would I do that?

Basically, I am trying to build a standard that acts as an envelope
around complete documents which can be of various different text-based
formats.

 

[usenet]

If I use the 'any' element in my schema to allow elements from another
schema to be used in instance documents based on my schema, is there a
way to force that the contents of that element must be an entire,
complete instance document for that other schema?

Let's say I have the following schema:

<?xml version="1.0" encoding="UTF-8"?>
<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema">
<xs:element name="BogusElement" type="BogusType"/>
<xs:complexType name="BogusType">
<xs:sequence>
<xs:any namespace="http://www.w3.org/1999/xhtml"/>
</xs:sequence>
</xs:complexType>
</xs:schema>

This would allow any global element from www.w3.org/1999/xhtml/.

<?xml version="1.0" encoding="UTF-8"?>
<BogusElement
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:noNamespaceSchemaLocation="Untitled1.xsd"
xmlns:xhtml="http://www.w3.org/1999/xhtml">
<xhtml:strong>Bolded Text.</xhtml:strong>
</BogusElement>

be a valid document under the above schema?
Yes.

How can I make it so that the contents of the BogusElement element in the
instance document must be a complete XHTML document in order to be valid?
Or would I just have to make that a verbal rule, not expressed in the
schema?

You could try the following schema approach...

<?xml version="1.0" encoding="UTF-8"?>
<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema">
<xs:element name="BogusElement" type="BogusType"/>
<xs:complexType name="BogusType">
<xs:sequence>
<xs:element xmlns:xhtml="http://www.w3.org/1999/xhtml"
ref="xhtml:xthml"/>
</xs:sequence>
</xs:complexType>

To complicate things further: Let's say I wanted that element to contain
only one of a list of valid, complete documents such as either a complete
html document OR a complete XHTML document OR a complete DocBook
document, all with headers. How would I do that?

Including an HTML document obviously has the problem that an HTML
document is not typically well-formed XML. But we can leave that
aside for the moment!

Also, I'm not sure of the valid set of DocBook schemas.

Given that, your schema could look something like:

<?xml version="1.0" encoding="UTF-8"?>
<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema">
<xs:element name="BogusElement" type="BogusType"/>
<xs:complexType name="BogusType">
<xs:choice> <!--Changed from sequence-->
<xs:element xmlns:xhtml="http://www.w3.org/1999/xhtml"
ref="xhtml:xthml"/>
<xs:element xmlns:docbook="http://docbook"
ref="docbook:docbook"/>
</xs:choice>
</xs:complexType>
</xs:schema>

HTH,

Pete.
--
=============================================
Pete Cordell
Codalogic Ltd
for XML Schema to C++ data binding visit
http://www.codalogic.com/lmx/
=============================================

 

[Boris Kolpackov]

Hi Grant,

If I use the 'any' element in my schema to allow elements from another
schema to be used in instance documents based on my schema, is there a
way to force that the contents of that element must be an entire,
complete instance document for that other schema?

If the schema for the namespace that you allow in your wildcard only
defines one global element that represents a complete document of
that vocabulary, then yes.

Let's say I have the following schema:

<?xml version="1.0" encoding="UTF-8"?>
<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema">
<xs:element name="BogusElement" type="BogusType"/>
<xs:complexType name="BogusType">
<xs:sequence>
<xs:any namespace="http://www.w3.org/1999/xhtml"/>
</xs:sequence>
</xs:complexType>
</xs:schema>

It has only the one element BogusElement and it can contain only the one
child element from the XHTML schema. Does this force the BogusElement
element in the instance document to contain a complete XHTML document
including headers or would

<?xml version="1.0" encoding="UTF-8"?>
<BogusElement
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:noNamespaceSchemaLocation="Untitled1.xsd"
xmlns:xhtml="http://www.w3.org/1999/xhtml">
<xhtml:strong>Bolded Text.</xhtml:strong>
</BogusElement>

be a valid document under the above schema?

It depends on the schema that defines the XHTML vocabulary. If it defines
strong as a global element then the above instance will be valid.

To complicate things further: Let's say I wanted that element to contain
only one of a list of valid, complete documents such as either a complete
html document OR a complete XHTML document OR a complete DocBook
document, all with headers. How would I do that?

I think if you have a predefined set of "content documents" then using
the choice construct might be a good idea. Alternatively, you can
introduce another level of indirection. With this approach you change
BogusType to read like so:

<xs:complexType name="BogusType">
<xs:sequence>
<xs:any namespace="http://www.example.com/envelop-content"/>
</xs:sequence>
</xs:complexType>

In other words, you only allow elements in place of a wildcard to be
from a namespace (or namespaces) that you control. Then you define
the envelop-content vocabulary like this:

<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema"
xmlns:xhtml="http://www.w3.org/1999/xhtml"
targetNamespace="http://www.example.com/envelop-content">

<xs:import ... import XHTML schema ...>

<xs:element name="xhtml-content">
<xs:complexType>
<xs:sequence>
<xs:element ref="xhtml::html"/>
</xs:sequence>
</xs:complexType>
</xs:element>

... Other *-content elements

</xs:schema>


hth,
-boris

 

[Grant Robertson]

<xs:element xmlns:xhtml="http://www.w3.org/1999/xhtml" ref="xhtml:xthml"/>

It looks like what you have done here is replace the

<xs:any namespace="http://www.w3.org/1999/xhtml"/>

component of my schema with an element-type component that lists a
namespace and then references the root element of that namespace. Is my
interpretation of this correct? Secondly, is this a valid XSD component?
What is this technique called (so I can look it up)?

 

[Grant Robertson]

Perhaps I should ask the more general question:

When people are creating an XML schema and they want to allow content
creators to insert full XML documents based on other schemas directly
within only certain elements of XML documents based on their new schema:
what do they do?

 

[Joseph Kesselman]

When people are creating an XML schema and they want to allow content
creators to insert full XML documents based on other schemas directly
within only certain elements of XML documents based on their new schema:
what do they do?

If you don't know which other schemas, make the content model of those
specific elements accept elements from other schemas (by using xsd:any).
I don't know of any way to say "accept only top-level elements/complete
documents"; once you open the window, anything can fly in.

Or you can spell out the specific root nodes you're willing to accept as
part of this element's content.

For an example of this, look at how the W3C has been designing XHTML to
allow plugging in fragments written in other specifications.

 

[usenet]

It looks like what you have done here is replace the

<xs:any namespace="http://www.w3.org/1999/xhtml"/>

component of my schema with an element-type component that lists a
namespace and then references the root element of that namespace. Is my
interpretation of this correct? Secondly, is this a valid XSD component?
What is this technique called (so I can look it up)?

That's right. Basically, it's a regular <xs:element ref=".../> schema
statement. The specification of the xhtml namespace prefix is
included in the element definition as it seems a convenient place to
do it, but it could equally have been defined in any of the parent
elements.

(Incidently - the current draft version of XSD 1.1 has better control
of xs:any wildcards. I wouldn't wait until the standard is published
and tools are available though!)

HTH,

Pete.
--
=============================================
Pete Cordell
Codalogic Ltd
for XML Schema to C++ data binding visit
http://www.codalogic.com/lmx/
=============================================