XML Validation: XML Schema (xsd) vs. Relax NG (rng) : Willing wheels

At first, I learned XML Schema but found it hard to deal with unordered xml nodes, plus it’s not that easy to read.

I found out Relax NG, read the tutorial and well, it does the job pretty well!

its syntax is self-explanatory
it allows to do anything you can do with XML Schema, with greater flexibility (cf. you can load external file, define and extend matching-classes…)
it’s far easier to read and memorize
you can transparently deal with unordered xml nodes (cf. <interleave> node)

It looks like we can validate xml docs with both XML Schema and Relax-NG using php, at least from version 5.2 (cf. XMLReader class on php documentation) and using DOM from svn snapshot (v5.3)

Validators

If you want to validate your xml file agains Relax-NG (or XML Schema or DTD), good news for linux users! You can use xmllint (man xmllint), which is available from libxml2-utils package (“sudo apt-get install libxml2-utils”).

Then run “xmllint –noout –relaxng schema.rng file.xml”.

misc

Below is the XML Schema I started to write:

<?xml version="1.0" encoding="UTF-8" ?>
<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema">
<!-- <query> "name" attribute pattern
		name is used as uri proxy => of the form: /[module]/[action] 	-->
<xs:simpleType name="xsdQueryName">
<xs:restriction base="xs:string">
<xs:pattern value="/[a-zA-Z]+/[a-zA-Z]([a-zA-Z]|-)+"/>
</xs:restriction>
</xs:simpleType>
<!-- <contact> in add-contact query -->
<xs:complexType name="xsdContactToAdd">
<xs:all><!-- this type defaults minOccurs="1" and maxOccurs="1" -->
<xsl:choice><!-- "user_fk" is provided when adding a new contact while "id" is provided when updating an existing contact -->
<xs:element name="id" type="xs:positiveInteger"/>
<xs:element name="user_fk" type="xs:positiveInteger"/>
</xsl:choice>
<xs:element name="email_address" type="xs:normalizedString"/>
<xs:element name="first_name" type="xs:normalizedString" minOccurs="0"/>
<xs:element name="last_name" type="xs:normalizedString" minOccurs="0"/>
<xs:element name="company" type="xs:normalizedString" minOccurs="0"/>
</xs:all>
</xs:complexType>
<!-- <contact> node -->
<xs:element name="contact" type="xsdContactToAdd"/>
<!-- <query> node -->
<xs:complexType name="xsdQuery" mixed="true">
<xs:sequence>
<xs:element name="contact"/>
</xs:sequence>
<xs:attribute name="name" type="xsdQueryName" use="required"/>
</xs:complexType>
<xs:element name="query" type="xsdQuery"/>
</xs:schema>

Not crystal clear, even though I used comments and split validations into sub-units to have a clearer doc

sources

http://www.oasis-open.org/committees/relax-ng/tutorial.html Relax-NG official tutorial
http://www.w3schools.com/schema/default.asp Learn XML Schema online with w3schools.com
http://es.php.net/manual/en/xmlreader.setrelaxngschemasource.php
http://es.php.net/manual/en/domdocument.relaxngvalidatesource.php
http://www.validome.org/xml/validate/ online xml validator, accept DTD and XML Schema as validation rules
http://www.w3.org/TR/xmlschema-1/ Official w3c XML Schema Technical Report
http://www.oasis-open.org/committees/relax-ng/spec-20011203.html Relax-NG official specs.
http://www.webreference.com/xml/column59/ XML Schema vs. Relax-NG quick (biaised) comparison
http://infohost.nmt.edu/tcc/help/xml/lint.html

Willing wheels