Abstract by Gio Wiederhold.
XML briefly
We describe only a few basic commands of the eXtended Markup Language
(XML). The current common version is XML 1.0.
XML formatted data are intended for processing, not just browsing.
Browsing of XML is currently supported mainly by Internet Explorer 5.0.
Conventions
XML is an application conforming to a W3 group standard. XML
uses embedded tags to indicate semantic structures, while leaving the
interpretation of the tags to the client's application programs. The
tags directives are bracketed by Less-Than (<) and Greater-Than
(>) symbols. To enable this note to be printed in an HTML format we
entered these symbols internally as (< for < and
> for >). XML also uses square brackets.
The characterset for XML is Unicode, a standard covering the
characters used for HTML, but also other alphabets. In fundamental
form they start with an ampersand (&), are denoted by an integer,
and end with a semicolon(;). However, the ASCII character code we use
is an accepted default subset.
General layout
Each document should start with a process instruction
<?xml version 1.0?>,
here indicating that the document conforms to XML version 1.0, followed by
the tagged root-name of the document.
A simple document showing an XML document with 4 early
Hitchcock movies
(requires Microsoft Internet Explorer 4.+) uses tags suitable
for movies, its body is hence called <movies>.
All commands have a corresponding closure, for instance there should
be a and tag </movies> at the end of the document.
If the contents is empty, an abbreviation is allowed combining
both tags: <no-content/>.
The only content permitted in XML are character strings (called
CDATA). No quotes are needed for CDATA. An example of a large
document, publicly available, is Shakespeare's
Taming of the Shrew. Internet Explorer will simply show it with
all the markups, since we have not defined a conversion for
presentation for it. But you can see how how the meangful tags now
allow searching for items as the characters (<PERSONA>) in the
play. All the tags used are listed in the accompaning
Data Description.
The Processing Schema -- DTD
It is important to document what elements can appear in an XML
document, and how they are to be arranged. That Notation is called
a Data Type Definition (DTD). We show
a sample DTD
used for plays of Shakespeare. It uses some of symbols (*,+,?) encountered
when presenting
regular expressions in the notes.
Notice that the tags are now related to the subject matter, and semantically
meaningful for people who understand plays. However, <FM> and <P+> are
mysteries until one looks at the contents; they denote 1 or more lines of
comments entered by the people who created the XML version of the play.
Looking at an XML document
Since XML does not specify how documents are to be presented,
that task is left up to the applications that read XML files. Without
any program, our browsers (currently Microsoft IE explorer 4.0 or
better 5.0) simply show the XML source
document, converted manually to an html representation (a list of
4 Hitchcock movies) as a simple list, with all the tags explicitly
shown. ( See list of Hitchcock
movies, in XML form).
To obtain well-formatted visible output XML data must be
converted to HTML, but should be done automatically. Such a
conversion could be done for any specific XML file by any program that
understands the tags, and accordingly creates a suitable HTML file.
Such a program can be written in JAVA, and that allows its execution
on the client computer (see Notes
for meeting 3). Such a JAVA applet is available in the Microsoft
Internet Explorer as XMLDSO (XML Data Source Objects). You can see a
simple example of it's use on the
4 Hitchcock and on the
full list can be seen. Such a longish list gets to be awkward.
For more detail, see
Microsoft Data Binding Documentation (Jan.2000).
To create prettier outputs, we have to generate fancier HTML
output. For instance, to split the long table for all of
Hitchcock's films by category we defined a table in the program.
We can also rearrange fields.
Long tables take long to load, and are hard to manipulate.
We can limit the size if the table to be shown with a
<TABLE DATAPAGESIZE=8 ID=table WIDTH=100% datasrc=#xmldso>
specification. To allow manipulation of that table we add a provision
for
<INPUT TYPE="button" VALUE="Next" ONCLICK="table.nextPage();">
of a button click, which refers to that table's ID. Now we can look at
Hitchcock's films page by page. This example shows all Hitchcock
movies, but not the directors heading (file Hitch0.xml).
Testing to include the directors heading
HTML 2-level table constructed by hand;
(file Hitch.xml) for Hitchcock's films page by page.
This formatting is created by the XML client application, here the
HTML program we have written. An alternate choice is to provide
formatting by a server, a style sheet can be specified.
XSL style sheets
With a suitable style sheet fancy
formatting can be generated.
Style sheets were common in SGML, so that a publisher could determine
the printing layout of a book. For XML file W3 has proposed a general
meta language, the XML Stylesheet Language
(XSL).
All tags that match the
XSL specification, are then not shown.
Examples show a formatted table with
four
Hitchcock movies as well as
all
Hitchcock movies, formatted in multiple tables, using an
XSL style sheet
source; The XSL
source
has been converted to HTML for readability under any browser.
An XSL interpreter in included in Microsoft IE version 5.0. If
no XSL is indicated, it will use a default style sheet, that is
book-oriented. It recognizes the following tags:
book?
(book
format).
In a stylesheet layout styles, relative sizes, and colors can be indicated
Cross References
The ability to go to other documents exists also in XML, as in HTML,
but differs in format.
XML Checkers
Notes
Contents of: Charles F. Goldfarb and Oaul Prescod: The XML Handbook,
3rd edition Prentice Hall, 2001.
XSL information
[W3C98]W3C: Extensible markup language. 1998.
[McGrath98] Sean McGrath: XML by Example: Building E-Commerce Applications; Prentice-Hall, Charles F. Goldfarb Series on Open Information Management, 1998.
See also the CS99I references.
For limitations see Madnick paper.