xml-dev - RE: how far should you go with XML schema?

["David Valera" <dvalera@pcl-hage.nl>]

> About your data, I would do this, using natural containment
> semantics. From a linguistic point of vue, it can be
> questionned to use a <wedding/> tag alone to indicate married
> status, but it avoids
> need for "if" construct in Schema.
> Also this design, using natural containment and tags from
> natural language like "wedding" and "date", is extendible.
> You can add later inside wedding, without breaking code,
> other tags such as <location>, <witness>, <contract>, etc.
>
> <xsd:element name="myroot">
>  <xsd:complexType content="elementOnly">
>   <xsd:element name="wedding">
>    <xsd:complexType>
>     <xsd:element name="date" type="xsd:date" minOccurs="0" />
>    </xsd:complexType>
>   </xsd:element>
>  </xsd:complexType>
> </xsd:element>


Well, in this case that would be a better way of implementing it. But this
was just an example of the problem I am facing. For example, I have an
element of type payment, wich can have a few different values:

	<xsd:simpleType name="payment" base="xsd:string">
		<xsd:enumeration value="each month"/>
		<xsd:enumeration value="each 2 months"/>
		<xsd:enumeration value="half year"/>
		<xsd:enumeration value="once a year"/>
		<xsd:enumeration value="at once"/>
	</xsd:simpleType>

For each of the values that payment can have here I would like to have some
restrictions on the elements that appear somewhere else in the XML file. A
good implementation of the structure of the XML file would solve most of the
problems like you described, but not all of them I am afraid.

I think I will take a look at schematron from Rick Jelliffe, since that
comes the closest to what I want to accomplish.

David Valera


***************************************************************************
This is xml-dev, the mailing list for XML developers.
To unsubscribe, mailto:majordomo@xml.org&BODY=unsubscribe%20xml-dev
List archives are available at http://xml.org/archives/xml-dev/
***************************************************************************