Using Metadata to create information services
So far this semester we have explored how metadata and structured information
is used as a core component of digital documents (e.g. HTML) and how
representations of print and digital documents can be described using multiple
metadata schemas (e.g. MARC, Dublin Core and Qualified Dublin Core) and
encoded using multiple encoding platforms (e.g. HTML, XML and MARC). As
our last task in exploring the creation and representation of digital documents we
are going to learn about the role of metadata and structured information in
creating information services.
For our purposes an information service is defined as “an operation or task that
fills a specific information need.” We experimented with information services
earlier in the semester when we used Cascading Style Sheets (CSS) to create
custom views of an HTML document. We have also seen how a suite of
information services can be designed together to help librarians accomplish the
cataloging tasks (e.g. MARCEdit).
In this exercise we are going to explore a technology called eXtensible
Stylesheet Language (XSL). XSL is similar to the other metadata schemas that
we have explored (HTML, DC, MARC) in that it has a defined set of elements
and attributes. It is also similar to HTML in that it is designed to be processed by
a specialized program. While HTML documents are designed to be processed
by web-browsers, XSL documents are designed to be processed by an XSL
processing engine. HTML Browsers have pretty good XSL processors in them
and we will see that in action but there are also specialized applications such as
Exchanger that allow you to apply an XSL style sheet to an XML file.
In this set of exercises, we will create XSL programs (also called XSL templates)
and become familiar with how to apply them to XML files. In our first exercise,
we will create a common ‘first program’ for many people, the hello world
program. In our second and third exercises, we will use an RSS/DC xml file and
see how XSL interacts with that file as we create HTML documents.. Our
development platform will be Exchanger. We will use a web-browser as our
transformation engine. Our goal for this exercise is to be familiar enough with the
transformation process to complete exercise 2
Exercise 1 – XML & XSL
In this exercise, we will create an XML file that contains the words “hello world”
and use XSL to retrieve data from that file.
Step 1:
Create a new XML file in your XML editor. This file follows a simple
made up standard using root as a main element and message as :
1.
<?xml version="1.0" encoding="UTF-8"?>
2.
<root>
3.
4.
<message>Hello World</message>
</root>
a. Save the XML file, giving it a name of your choice
In the above XML file take a moment to identify the core elements:
The root element of the document is:
The value of the element <message> is:
Recall for a moment the work we did with the Document Object Model (DOM).
The DOM defines a method that we can use to access elements and attributes in
our document and help us work with documents to either control display,
manipulate content or, in this case, run a program that turns that content into a
new document.
XSL traverses the DOM using a method called XPath. XPath is a querying
language similar in many ways to structured query language (SQL) used in
relational database management systems. Just like SQL, XPath allows you to
get specific data from your XML document. XPath statements are pretty easy to
create:
XPath statement to get the value of the <message> element in our document
5.
/root/message
Pretty underwhelming right? XPath in essence looks a lot like a hierarchical
filesystem structure because it basically is a hierarchical structure. XPath
supports absolute references (e.g. /root/message) and contextual references
(e.g. ./message). A contextual reference is usually used within the context of
another element (more about this later!).
In order to get the value of our <message> element we use an XSL element
called <value-of>
6.
<xsl:value-of select=”/root/message”/>
Key Questions
Take a moment to examine the code here and answer the following questions
What is the namespace of the value-of element?
What attribute does that element have?
What is the value of the attribute of your element?
What is the meaning of the last “/” in the <xsl:value-of…> element?
Incredibly, when you apply an XSL stylesheet to an XML document it pulls the
data from the XML document and displays it according to the instructions in your
XSL resource.
Step 2:
Now that we have a basic understanding of how XSL and XPath work
Let’s create an XSL file that we will use to transform our XML file into
an HTML file.
a. Create an XSL file in your XML editor:
b. Click on New >> select XML file. If prompted, choose “1.0”
Notice the declaration of the XSL namespace in the stylesheet element. This is
just like our work with Dublin Core in XML. You have to define the namespaces
that you are going to use in your document. The powerful aspect here is that
defining a namespace brings in a suite of functions that we will use to process
our XML document! Notice also the use of HTML elements in the program. XSL
documents are primarily designed for transformation and output – they access
values in a source XML document and echo those values along with formatting to
the screen or to a new document.
Key Questions
Take a moment to look at the program below and answer the following questions:
7.
8.
<xsl:stylesheet version="1.0"
xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
<xsl:template match="/">
9.
<html>
10.
<body>
11.
<h1>
12.
<xsl:value-of select="/root/message"/>
13.
</h1>
14.
</body>
15.
</html>
16.
</xsl:template>
17.
</xsl:stylesheet>
Look at the element <xsl:template> Notice that this element has an
attribute (match) with the XPath value (/). Assume that we called this
stylesheet from our other XML document. Where in this source XML
document does this XPath expression match?
Consider what you have learned so far about XSL. If we used this XSL file
to transform our XML document what type of document would be output?
Step 3:
Type the program into Exchanger and save it. Make sure your XSL
stylesheet is well formed and validates.
Step 4:
Lets transform our XML document using our XSL stylesheet. Assign
your XSL file to your XML file
a. Put the following line of code in the top of your XML document right
under the XML element declaration and before your <root>
element.
18.
<?xml-stylesheet type="text/xsl" href="[XSLFILENAME.xsl]"?>
This command links our XSL stylesheet to our XML document. The two
key attributes here are type (defines the type of document) and href
(defines the location of the document)
Step 5:
From within a web-browser, Open your XML file. What do you see?
Key Questions
Consider the following schematic representing a webpage built with XML, XSL
and CSS.
What would some advantages and disadvantages of this approach be as
compared to the single document approach that we worked with earlier in
the semester?”
Advantages
Disadvantages
Exercise 2 – Working with multiple records in an XML file
In our first exercise we found that XSL can transform an XML document into an
HTML document. XSL also includes a set of operations that allow us to work
with multiple records in a single file. In fact XSL has so many functions that it ca
be used to create a dynamic web service! In this exercise, we will get started
with this by working with an RSS file of Dublin core records coded in XML.
Step 6:
During this exercise, we will use XSL to create a list of the titles in the
XML file coded in HTML.
a. Download the XML file from our course website
(http://erikmitchell.info/lbsc_670/class_07_xsl0) for this exercise
(Newbooks.xml)
b. In this exercise, we want to loop through each title in our RSS file
and output the title with a link to the record in the catalog. First,
let’s look at the RSS file. Open the newbooks.xml file and answer
the following questions:
Key Questions
What is an absolute XPath expression to the <item> element?
What is an absolute XPath expression to the <title> element of the RSS
feed?
What are the relative (e.g. from <item>) XPath expressions for the book title
and link?
Book XPath:
Link XPath:
Before we move too far along with this process we need to understand one more
XSL element
<xsl:for-each>: This element creates a looping structure that repeats for every
instance of an element. So in our XML file we have multiple <item> elements.
This means that we will use an <xsl:for-each> loop to get to every <item>
element.
With this information in hand (XPath expressions for <title>, <item> and relative
expressions for book title and link) and our understanding of <xsl:for-each> lets
examine our XSL program:
19.
<xsl:stylesheet version="1.0"
xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
xmlns:dc="http://purl.org/dc/elements/1.1/"
xmlns:dcterms="http://purl.org/dc/terms">
20.
<xsl:template match="/">
21.
<html>
22.
<body>
23.
<h1>
24.
<xsl:value-of select="/rss/channel/title"/>
25.
</h1>
26.
<ul>
27.
<xsl:for-each select="/rss/channel/item">
28.
<li>
29.
<xsl:value-of select="./dc:title"/>
30.
</li>
31.
</xsl:for-each>
32.
</ul>
33.
</body>
34.
</html>
35.
</xsl:template>
36.
</xsl:stylesheet>
Key Questions
Review the code above and answer the following questions. Lets pay attention
to a specific section of code (look below). This code executes the <xsl:for-each>
process which loops through each instance of the <item> element in our RSS
file.
1.
<xsl:for-each select="/rss/channel/item">
2.
<li>
3.
<xsl:value-of select="./dc:title"/>
4.
</li>
5.
</xsl:for-each>
Note how within the <xsl:for-each> element we define a generic process
(e.g. output an li element, output a link to a feed item, and close the li
element). This will be repeated for every <item> element in the file. Take a
moment to write out the HTML for the first instance of <item> in your XML
file based on the instructions in this set of code
Step 7:
Let’s continue creating our XSL file. Create a new XSL file in your
XML editor (see above instructions)
Step 8:
Code your XSL file to:
a. Output the name of the feed in an <h1> element
b. Go into each item element and output the title of the book
c. Enter the code from our example 2 xsl file
d. Save the file
Step 9:
Associate the XSL file with the newbooks.xml document (See previous
instructions in this worksheet)
Step 10: Load the newbooks.xml document in your webbrowser. What
happened?
Exercise 3 – Adding sorting and links to our list
In this exercise, we will sort our results and add links to our list of titles. In order
to do this, we will have to use the <xsl:attribute> element. The code below is
substantially the same as our above example with the exception of the addition of
the <a> element (see bolded code)
A few notes on the new code:
 Notice how we use <xsl:attribute name=””/> to define an attribute for our
<a> element

We will use the <xsl:sort select=””/> element to have the file sort by title
1.
<xsl:stylesheet version="1.0"
xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
xmlns:dc="http://purl.org/dc/elements/1.1/"
xmlns:dcterms="http://purl.org/dc/terms">
2.
<xsl:template match="/">
3.
<html>
4.
<body>
5.
<h1>
6.
<xsl:value-of select="/rss/channel/title"/>
7.
</h1>
8.
<ul>
9.
<xsl:for-each select="/rss/channel/item">
10.
<xsl:sort select="./dc:title"/>
11.
<li>
12.
<a><xsl:attribute name="href"><xsl:value-of
select="./link"/></xsl:attribute><xsl:value-of select="./dc:title"/></a>
13.
</li>
14.
</xsl:for-each>
15.
</ul>
16.
</body>
17.
</html>
18.
</xsl:template>
19.
</xsl:stylesheet>
Step 11: Implement the changes to your XSL file. Save your XSL file and
refresh newbooks.xml in your web-browser. Did you get links? Did the
items sort properly?
Wrap up
This exercise has been our first foray into programming and software design. If
you master the exercises in this worksheet you have the skills required to
complete assignment 2. As the semester continues we will explore other forms
of structured information. Each of these structures (e.g. Taxonomies, SOLR
indexes) will introduce us to different aspects of the relationship between
structured data and information services. There is a wide range of systems and
programming platforms that we are not going to talk about this semester though.
If you enjoyed your experience with XSL I would recommend continuing down
that path and learning more about XSL systems and even a full-fledged
programming language like Python, PHP or Perl.