XML

eXtensible Markup Language
Key technology for future Web applications
Applications of XML
- New publishing format
- Application integration

New Publishing Format

Initially conceived as a new publishing format to replace HTML
Logical model of data from which different types of publications could be generated
Example of this model
- WSJ Interactive uses XML to mark up its articles

Motivation

Using XML as a publishing format was motivated by the significant investments businesses had made already in putting their data on-line
Goal of businesses was to maximize their use of this data by being able to re-purpose it to different media
- Publish to the web, print, audio
But as the understanding of XML technology grew, it became clear that its real power is elsewhere ...

Application Integration

XML offers a powerful new way to integrate new and existing applications
A powerful idea: Replace N x N translators between applications by a common data format
Tags on an electronic invoice could identify some data as a price and other data as a quantity
Problem remains of getting application developers to agree on what those tags mean

Promise

XML holds a huge promise for enterprise application integration (EAI)
- You can add code to applications that emits messages as XML ,and to other applications so it receives and processes XML data
XML is thus at the heart of B2B commerce
- XML allows communication with XML-aware applications in another enterprise

XML as Middleware

XML - Problem

Have you ever tried writing a parser to extract data from a web page (in HTML), e.g. a stock quote?

<table>
  <tr>
    <th>Symbol</th>
    <th>Price</th>
    <th>Volume</th>
  </tr>
  <tr>
    <td>YHOO</td>
    <td><b>412</b></td>
    <td>5000000</td>
  </tr>
</table>

XML - Problem

If you thought that was easy
- What about maintaining your code?
Let's be specific and consider what you would happen, if some day the web master of your stock quote site decides to add a banner ad at the top of the page
- She puts the banner ad in a table by itself
- As the table with the stock quote becomes the second table, your parser fails!
Your parsing code fails with potentially ludicrous results!

XML - Solution

XML steps to the rescue
- The X in XML says XML is extensible
- XML allows you to define your own tags and the semantics for them

Our stock quote example could be rewritten in XML

<quote>
    <symbol>YHOO</symbol>
    <price>412</price>
    <volume>5000000</volume>
</quote>

XML - Solution

The key idea of this approach is to
- SEPARATE DATA FROM ITS PRESENTATION
Ensuing benefits
- Same data can be presented in multiple views and used for different purposes
- Multiple data views are easier to maintain (to publish changes, simply re-generate the views)

XML and HTML

Both XML and HTML are markup languages
Common ancestor from SGML
Both use tags for marking up data
However, the tag set of HTML is fixed
HTML focuses on formatting only

How Does XML Work?

Parse the Data ...

Data publishers encode their data in XML
An XML parser in the application parses the document (optionally using a DTD)

Document Types

The syntax is defined in a Document Type Definition (DTD) specified in the document
- A grammar which defines the rules the document has to follow (e.g. nesting of elements)
For example, a quote consists of a stock symbol, its current price and trading volume
```
<!ELEMENT quote (symbol, price, volume)>
```

... display, ...

Developers write stylesheets that define the rendition of the document on multiple devices, or for multiple users
- Different devices (browsers, handhelds, phones)
- Different users (manager, developers)
Transformation rules in XSL, or programmatically defined

XSL

eXtensible Stylesheet Language
Stylesheets define a set of transformation rules to be applied to the data in the XML document
Using stylesheets one and the same data can be presented in multiple output formats
Transformation rules are themselves represented in XML

... process, ...

Applications can parse an XML document, and process it by mapping a shared format to their internal formats

... or retrieve

Developers can query the XML document
Query results are then passed back to the originating application (in XML format)

Advantages of XML

Leverages from being a text-based protocol
XML provides a logical model of data, which can be presented in multiple different ways and used to share data between apps
Widely supported through standard tools
- XML parsers
- XSL stylesheet processors
- XML editors
- XML-based servers
Not based on proprietary technology

Application Examples

Wide range of applications for XML, such as
- Representing molecular structures
- Marking up web services for electronic commerce
- Defining webcast channels
- Exchanging data between enterprise applications
- Querying on-line catalogs
- Programming telephony applications

Creating Markup Languages With XML

XML is not simply a language, but a meta-language (a language for defining other languages)
Some examples
- CML (Chemical Markup Language)
- CDF (Channel Definition Format)
- WIDL (Web Interface Definition Language)
- RDF (Resource Definition Language)
- MathML (Mathematical Markup Language)
- WML (Wireless Markup Language)
- SOAP (Simple Object Access Protocol)

XML-related Technologies

XML Stylesheet Language (XSL)
XML Path Language (XPath)
Schemas and Datatypes (XML Schema)
XML Link Language (XLink)

Example

The following XML document contains the structure and contents of a book

<?xml version="1.0">
<!DOCTYPE book SYSTEM "book.dtd">
<book>
    <chapter>
        <par author="mw">First paragraph</par>
        <figure file="fig1.gif">
            <caption>The first figure</caption>
        </figure>
        <par author="cs">Second paragraph</par>
    </chapter>
    <chapter>
        <par author="mw">Chapter 2, first paragraph</par>
        <example>
            <caption>The first example</caption>
        </example>
    </chapter>
</book>

Example Explained

Processing instruction indicating the XML version
```
<?xml version="1.0">
```
Document type declaration (an XML document is self-describing in that it defines its grammar)
```
<!DOCTYPE book SYSTEM "book.dtd">
```
This header is followed by the root tag
```
<book>
```

Example Explained

There is only one root tag per document, which can have as many children elements as allowed by the DTD
A book contains chapters
```
    <chapter>
```
Each of the chapters contains paragraphs, figures and examples (in any order)

Example Explained

A paragraph specifies its author, otherwise it contains no other elements, just text
```
        <par author="mw">First paragraph</par>
```

A figure specifies the file name of an image and a caption in a child element

        <figure file="fig1.gif">
            <caption>The first figure</caption>
        </figure>

Example Explained

The first chapter contains another paragraph

        <par author="cs">Second paragraph</par>

before it closes (every tag has an end tag)
```
    </chapter>
```

This book has one more chapter comprising one paragraph and an example

    <chapter>
        <par author="mw">Chapter 2, first paragraph</par>
        <example>
            <caption>The first example</caption>
        </example>
    </chapter>

Finally, the book comes to an end
```
</book>
```

Document Object Model

Object representation of the XML document

Document Object Model

The document object model contains
- Element nodes
- Text nodes
As well as nodes (not shown in above diagram) representing
- Attributes
- Document itself
- Processing instructions
- Document type declaration
- Entities
- Entity references
- Comments

Document Type Definition

Restricts the structure of a class of XML documents
A DTD specifies
- which components are available for a particular type of document
- how these components can be mixed in order to produce a valid document
For example
- A book must contain at least one chapter
- A chapter can contain paragraphs, figures and examples (in any order)
- A chapter must include at least one of these
Syntax for declaring the document type
```
<!DOCTYPE root SYSTEM file>
```

Document Type Definition

Here is a DTD for the book example, which we will dissect shortly

<!ELEMENT book (chapter+)>
<!ELEMENT chapter (par|figure|example)+>
<!ELEMENT par (#PCDATA|emph)*>
<!ATTLIST par author CDATA #IMPLIED>
<!ELEMENT figure (caption)>
<!ATTLIST figure file CDATA #REQUIRED>
<!ELEMENT example (caption)>
<!ELEMENT caption (#PCDATA)>

<!DOCTYPE book SYSTEM "book.dtd">

Element Declarations

Documents are mostly composed of elements
Element declarations specify what kind of content a particular element may have
```
<!ELEMENT name content>
```
For example, the following expresses that a book must contain at least one chapter
```
<!ELEMENT book (chapter+)>
```

Content Model

Character data
```
<!ELEMENT caption (#PCDATA)>
```

Sequence

<!ELEMENT card (name, address, phone, email?)>

Repetition
- * (zero or more)
- + (one or more)
- ? (zero or one)
```
<!ELEMENT par (#PCDATA|emph)*>
```

Choice

<!ELEMENT chapter (par|figure|example)+>

Attribute List Declarations

Attribute lists define which attributes may be associated with a particular element

<!ATTLIST element_name
    attribute_name attribute_type default_declaration
    ...
    attribute_name attribute_type default_declaration>

Attribute List Declarations

Required (must be supplied)
```
<!ATTLIST figure file CDATA #REQUIRED>
```
Implied (filled in, if not supplied)
```
<!ATTLIST par author CDATA #IMPLIED>
```

Implied with default value

<!ATTLIST quiz answer CDATA #IMPLIED "yes">

Fixed (fixed value, cannot override)

<!ATTLIST book language CDATA #FIXED "en">

Attribute Types

String
```
CDATA
```
Enumerated
```
(com|net|org)
```

Tokenized

ID 		// unique element identifier
IDREF 		// reference to another element
NMTOKEN 	// strings without spaces

Example

DTD for a movie database

<!ELEMENT movies (movie*)>
<!ELEMENT movie (name, director)>
<!ATTLIST movie
    rating (G|PG|R) "R"
    duration NMTOKEN #REQUIRED>
<!ELEMENT director (#PCDATA)>

References

Chapter 14 in CGI Programming with Perl