Continued from XML Namespaces for Dummies – Part 1.
Namespaces also give us several brand new opportunities. You may be familiar with ISBN numbers. These numbers give a unique number to every book ever published. What if the Publications Consortium and the Bookshelf Alliance decide they want to include ISBN numbers for their file formats? They can use XML namespaces to announce to the world that the ISBN number schema is actually controlled by someone else. This feature of mixing and matching namespaces is very powerful. The following figure demonstrates how the publications file format might be updated for version 2.0:
Figure 7: Adding a second XML namespace isbn
<field name="title">Adventures of Huckleberry Finn</field>
<field name="author">Mark Twain</field>
<field name="fullTitle">The annotated Huckleberry Finn : Adventures of Huckleberry Finn (Tom Sawyer's comrade)</field>
There are a few things to take notice of here. The first is the peculiar definition of the ISBN namespace. Notice how the namespace name is qualified in the format xmlns:isbn. This special format signals that the namespace applies only to elements with the same prefix qualifier–In this case, isbn. Thus, the number element has the isbn prefix, meaning that it is defined by the schema associated with the namespace with that prefix. You can trace the definition back to where it was defined by traveling up the number element's ancestors up to the publications element.
The second thing to notice is that the XML namespace applies to everything within the element. We say that the scope of the XML namespace is the element and its descendent elements. Thus, we can define xmlns:isbn on the publications element, on the book element, or even on the number element, and it is still equivalent because it applies to everything within. Said another way, the number element is in the scope of the xmlns:isbn namespace no matter which of these it is defined on. We might also say that the publications element is the widest scope and that the number element is the narrowest scope. Conventionally, you define namespaces on the top level element because this makes them easiest to find.
Consider how the following are equivalent because of these two rules regarding prefixes and scope:
Figure 8: Equivalent due to prefix equality and namespace scopes
In the first case, we use the prefix i to define the namespace. This shows that we can use any prefix name that we like when defining a namespace. In the second instance, we define the XML namespace directly on the number element. In this second case, the namespace defined there overrides the namespace defined at a wider scope. That's another rule: namespaces apply to the narrowest scope. We will consider this momentarily when we override namespaces
Let's take a moment from the hows and whys of namespaces to consider how some previously mystifying ideas in XPath are all of a sudden quite mundane.
For instance, namespaces explain why XPath 2.0 functions work. For instance, let's say that we wanted to count the number of books in a bookshelf file:
Figure 9: Using the count function to count the number of books
The use of the fn: prefix in XPath 2.0 indicates that a namespace is being used and that the namespace belongs to the consortium originally responsible for XML and XPath. In some reporting tools and XML development environments, you can even define how your own functions but that's beyond the scope of this article.
The use of prefixes was also shown in Figure 6.
There are a few interesting rules when dealing with namespaces. The first is the notion of the default namespace. The original file defined in Figure 1 didn't declare an XML namespace. When no namespace is defined, the default namespace is in effect.
Explicit namespaces are supposed to have schemata, and these schemata can be enforced. Most of the time when you use an explicit XML namespace, if your document does not conform to the schema, it leads to an error being generated. On the other hand, the default namespace doesn't have a schema. More precisely, the default namespace has a schema that says anything goes. This explains why you can create basic XML documents with anything you want in them without errors being generated.
A second interesting idea is the question of attributes. You can also apply namespaces to an attribute by using the prefix notion.
Figure 10: Using attributes in a particular namespace
<books xmlns="http://schemas.bookshelfalliance.org/2011/bookshelf.xsd" xmlns:i="http://schemas.isbn.org/ns/1999/basic.dtd">
<title >Adventures of Huckleberry Finn</title>
Ordinarily, attributes belong to the default namespace. They only ever belong to an explicit namespace when xmlns= is used. For instance, in the publications schema, the name attribute on the field element belongs to the http://publicationsconsortium.org/schemas/publications_version_2.dtd namespace. However, when you use prefixes, the attribute belongs to the default namespace. So, in Figure 8, the attribute on the isbn:number element is not in the isbn namespace; it is in the default namespace.
A third advanced idea is the question of the implicit namespace. You might be wondering whether xmlns attribute is simply "magical". What namespace does it belong to? Why can the attribute xmlns:prefix be used? Technically, every XML document defines an XML namespace corresponding to the definition of XML itself. This namespace is implicit, so it is always defined. The xmlns attribute is defined in the schema of this particular namespace. In the case of xmlns:prefix, another implicit namespace defining the schema of XML Namespaces itself also exists. This accounts for the syntax and explains why xmlns is not actually magical.
A fourth and final advanced idea is the idea of reverting to the default namespace. If you define an explicit namespace such as in Figure 7, how can you revert to the default namespace? You do that by adding the attribute xmlns="". In XML Namespace 1.1, a small extension was added that also lets you also revert namespaces with prefixes. For instances xmlns:isbn="" would revert the namespace in Figure 7 back to the default namespace.
XML namespaces are for sharing. Namespaces can be used to mix and match schemata. When combined with XPath, new possibilities of complex and integrated XML files is opened up.