As you read in the previos chapter, XForms accomplishes the data capture through two components, XForms Model and XForms User Interface. The connection between these two parts is called binding, and it uses the common W3C technology called as XPath.

In order to take the advantages of binding, form authors need to provide the XML template, called as instance data, which provides the place to reside for entered data. The instance of data can either be empty initially (for the blank form) or can hold an initial data (in case of the pre-populated form). The ref attribute on each of the form control, seen previously, actually holds the XPath expression that points to the location in an instance the data.

If you've had some sort of experience with XPath, you probably know that the XPath deals with a node-sets, where a node is the fundamental bit of the XML such as an element or an attribute. Whenever the XForms uses ref attribute, it aplies the first node rule, so that even if the XPath would normally return several nodes, only the first (in the order that things appears in a document) is used by the XForms. In contrast, the attribute node set indicates that the multiple nodes are in play.

Any of the serious XML work will inevitably run into an XML namespaces, with a telltale xmlns attributes. Since this is the XForms Institute, and not a Namespaces Institute, several of the simplifications are used here:

The default namespace
A default namespace is applied to an XHTML. It is possible that by the time XHTML 2 is finialized that an XForms will share a namespace, so the XForms elements too are considered as default. .

The user namespace
A prefix my: is used for most of the user-provided instance data.

Others
The few other namespace prefixes, such as ev: for an XML Events may appear, and it will be explained as needed.

Keep in mind that in the XPath expressions, no default namespace is applied, and thus the prefixes should be used liberally.

First of all, as the child of a model element, an element called as instance is needed to provide an instance of the data, which can be either an inline XML, or in the separate document pointed to by the src attribute. Cosider for an example, a fragment of the UBL document:

This sample of code includes a submission element as before, and additionally an instance of the element populated with a real-world XML. It also has the bind element, the key to the power of an XForms: it can do work directly with nearly any kind of the XML in existence.

Others
A nodeset attribute gives the hint of how this element will work: it selects all of the my:Description elements in a document. In the purchase order, you would expect several of the line items to be present, each with the description field. An XPath expression do selects them all, and applies the property called "required" to each of them.

An XPath works much like the directory path, with each step descending one level into an XML. An attribute step is accomplished with the leading character @, as in an html:a/@href.