DTD

📎 Examples

Definition

XML application

fixed set of tags for a specified application / use case

DTD

= Document type definition

  • specification / schema language

    validates XML documents

    lists all elements and attributes of documents

  • Not in XML syntax

Limitations of DTD

  • No control over exact number of children
  • Limited data types
  • No inheritance
  • No support for namespaces

1) Element declaration

<!ELEMENT elementNamespecification>

Empty

<!ELEMENT elemEMPTY>

allows <elem></elem> or <elem/>

No restrictions

<!ELEMENT elemANY>

allows any child that was declared before

String only

<!ELEMENT name(#PCDATA)>

only parsed character data , no children

Fixed set of children

Only a one child per element, nothing else

parent:<!ELEMENT parent(child)>

child:<!ELEMENT child(#PCDATA)>

Only two child elements, nothing else

parent:<!ELEMENT person (first, last)>

child 1:<!ELEMENT first (#PCDATA)>

child 2:<!ELEMENT last (#PCDATA)>

Can also be used for enums

<!ELEMENT day (Mon | Tue | Wed)>

<!ELEMENT Mon (#PCDATA)>

<!ELEMENT Tue (#PCDATA)>

<!ELEMENT Wed (#PCDATA)>

Mixed content

<!ELEMENT definition(#PCDATA | child)*>

combination of text and elements

Setting number of children

Occurrence indicators ? , * , +

child? zero or one

child* zero or more

child+ one or more

Can be used for more complex expressions

Namespaces

The validator does not care about name spaces - they make no difference

  • Example
    <!ELEMENT course (title, assessment, univ:assessment)>
<!ATTLIST course xmlns CDATA #FIXED "www.oeh.ac.at">
<!ATTLIST course xmlns:univ CDATA #REQUIRED>
<!ELEMENT title (#PCDATA)>
<!ELEMENT assessment (#PCDATA)>
<!ELEMENT univ:assessment (#PCDATA)>
    <!-- Students' and University's Evaluation -->
<course
xmlns="www.oeh.ac.at"
xmlns:univ="www.tuwien.ac.at"><title> SSD </title>
<assessment> Fair </assessment >
<univ:assessment> Elective </univ:assessment >
</course>

2) Attribute declaration

<!ATTLIST elementName attributeNameattributeType attributeDefault>

Attribute types

(...|...) only one of the two given values allowed

CDATA any string

NMTOKEN name token = legal XML-names, that can start without letter, underscore

NMTOKENS list of nmtokens, seperated by spaces

ID must be unique (only used once in element)

IDREF must be ID from other element

IDREFS list of IDREFs, seperated by spaces

…read XML in a Nutshell, Chapter 3 for more details

Attribute defaults

#IMPLIED optional

#REQUIRED required

#FIXED value default value, assigned everywhere, can’t be changed

value default value, can be changed

3) Entity declaration

<!ENTITYshortcut "expanded name">

Entity declaration

We can add more to the 5 predefined XML entities: lt , gt , amp , quot , apos

ie.<!ENTITY ssd "Semi-structured Data">then we use&ssd;instead of typing it out

Declaring the DTD

Document type declaration (≠ DTD)

Referencing the DTD after the XML declaration

  1. Using URL or local path
    <!DOCTYPE person SYSTEM "http://www.mysite.com/dtds/person.dtd">
<!DOCTYPE person SYSTEM "/dtds/person.dtd">
<!DOCTYPE person SYSTEM "person.dtd"> <!--If in the same directory-->

    Keyword SYSTEM if it is user-defined, PUBLIC if its publically available 

    <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN" "xhtml11.dtd">

  1. Fully defining DTD in the same document
    <?xml version="1.0" encoding="UTF-8" standalone="yes"?>
<!DOCTYPE person [
<!ELEMENT person (name, tel, fax, email+)>
<!ELEMENT name (first, last)>
<!ELEMENT email (#PCDATA)>
]>
...

  1. Partially defining DTD in the same document

    DTD = internal DTD subset + external DTD subset

    <?xml version="1.0" encoding="UTF-8" standalone="no"?>
<!DOCTYPE person SYSTEM "person_text.dtd" [
<!ELEMENT person (name, tel, fax, email+)>
<!ELEMENT name (first, last)>
]>
...
    <!ELEMENT email (#PCDATA)>