Creating DTDs

Welcome to our free XML tutorial. This tutorial is based on Webucator's Introduction to XML Training course.

Contact Us or call 1-877-932-8228
Creating DTDs

Creating DTDs

DTDs are simple text files that can be created with any basic text editor. Although they look a little cryptic at first, they are not terribly complicated once you get used to them.

Note: we will review a DTD document in the next presentation.

A DTD outlines what elements can be in an XML document and the attributes and subelements that they can take. Let's start by taking a look at a complete DTD and then dissecting it.

Code Sample:

<!ELEMENT beatles (beatle+)>
<!ELEMENT beatle (name)>
<!ATTLIST beatle
	real (yes|no) "yes">
<!ELEMENT name (firstname, lastname)>
<!ELEMENT firstname (#PCDATA)>
<!ELEMENT lastname (#PCDATA)>

The Document Element

When creating a DTD, the first step is to define the document element.

<!ELEMENT beatles (beatle+)>

The element declaration above states that the beatles element must contain one or more beatle elements.

Child Elements

When defining child elements in DTDs, you can specify how many times those elements can appear by adding a modifier after the element name. If no modifier is added, the element must appear once and only once. The other options are shown in the table below.

Modifier Description
? Zero or one times.
+ One or more times.
* Zero or more times.

It is not possible to specify a range of times that an element may appear (e.g, 2-4 appearances).

Other Elements

The other elements are declared in the same way as the document element - with the <!ELEMENT> declaration. The Beatles DTD declares four additional elements.

Each beatle element must contain a child element name, which must appear once and only once.

<!ELEMENT beatle (name)>

Each name element must contain a firstname and lastname element, which each must appear once and only once and in that order.

<!ELEMENT name (firstname, lastname)>

Some elements contain only text. This is declared in a DTD as #PCDATA. PCDATA stands for parsed character data, meaning that the data will be parsed for XML tags and entities. The firstname and lastname elements contain only text.

<!ELEMENT firstname (#PCDATA)>
<!ELEMENT lastname (#PCDATA)>

Choice of Elements

It is also possible to indicate that one of several elements may appear as a child element. For example, the declaration below indicates that an img element may have a child element name or a child element id, but not both.

<!ELEMENT img (name|id)>

Empty Elements

Empty elements are declared as follows.


Mixed Content

Sometimes elements can have elements and text intermingled. For example, the following declaration is for a body element that may contain text in addition to any number of link and img elements.

<!ELEMENT body (#PCDATA | link | img)*>

Location of Modifier

The location of modifiers in a declaration is important. If the modifier is outside of a set of parentheses, it applies to the group; whereas, if the modifier is immediately next to an element name, it applies only to that element. The following examples illustrate.

In the example below, the body element can have any number of interspersed child link and img elements.

<!ELEMENT body (link | img)*>

In the example below, the body element can have any number of child link elements or any number of child img elements, but it cannot have both link and img elements.

<!ELEMENT body (link* | img*)>

In the example below, the body element can have any number of child link and img elements, but they must come in pairs, with the link element preceding the img element.

<!ELEMENT body (link, img)*>

In the example below, the body element can have any number of child link elements followed by any number of child img elements.

<!ELEMENT body (link*, img*)>

Using Parentheses for Complex Declarations

Element declarations can be more complex than the examples above. For example, you can specify that a person element either contains a single name element or a firstname and lastname element. To group elements, wrap them in parentheses as shown below.

<!ELEMENT person (name | (firstname,lastname))>

Declaring Attributes

Attributes are declared using the <!ATTLIST > declaration. The syntax is shown below.

<!ATTLIST ElementName
	AttributeName AttributeType State DefaultValue?
	AttributeName AttributeType State DefaultValue?>
  • ElementName is the name of the element taking the attributes.
  • AttributeName is the name of the attribute.
  • AttributeType is the type of data that the attribute value may hold. Although there are many types, the most common are CDATA (unparsed character data) and ID (a unique identifier). A list of options can also be given for the attribute type.
  • DefaultValue is the value of the attribute if it is not included in the element.
  • State can be one of three values: #REQUIRED, #FIXED (set value), and #IMPLIED (optional).

The beatle element has two possible attributes: link, which is optional and may contain any valid XML text, and real, which defaults to yes if it is not included.

<!ATTLIST beatle
	real (yes|no) "yes">

This tutorial is based on Webucator's Introduction to XML Training Course. We also offer many other XML Training courses. Sign up today to get help from a live instructor.