XML - An Introduction For System Administrators

• First let's recognize that HTML was in fact a good thing, for it's time (and is still a good thing for many uses -- like this presentation). There are, however, flaws that present obstacles to an information based society (pun intended).
• The data (information) and the formatting (how to display the data) are inseparable in an HTML document.
• There is no guarantee, even at that, that the document will display the same on different browsers.
• There is no "browser enforcement" of html standards, such as intermixing of tags.

• For starters, THE DATA AND FORMATTING ARE SEPARATED.
• What, you missed that?  Well, here's a little (incomplete) example in HTML and XML.

The <group> and <location> tags are defined elsewhere in a Document Type Definition (SSG_example.dtd).  The DTD's contain the description of all the tags that appear in a XML document.  DTD's are very important.  They define the structure of an XML document.

How does a parser parse the <group> tag, given we just made it up? Well, we didn't. it was defined in the DTD as an element, along with the attributes for the tag.

• So, now the browser (or XML-enabled application, possible embedded in a cell-phone, or more practically the server sending the HDML to the cell phone...) has the ability to discern what data to display and what not to display.
• Hmmm. XML-enabled? This means that the client-server model can be muddied up with intermediate servers acting as data-filters and data-gatherers.

• DTD - Document Type Definitions
• XSL - Extensible Style Sheets  for XML
• Well-formed, (they really mean it this time)
• Validated
• Schemas - 'Meta Data' now officially Document Definition Markup Language

• Elements - think of these as the individual "variables" of an XML document.  In our example <group> and <location> are elements.
• Attributes - these describe attributes of elements, the class, borderwidth, etc.
• Attribute List Declarations - describe the attributes an element will have.  May supply defaults, etc.
• Entities - these define specific content, as in an external file (xml or otherwise), or internal text for the document (as in <!ENTITY SSG  "Seattle Sage Group">  would later be referred to as &SSG; )

• Lark @ http://www.textuality.com/Lark (non-validating parser)
• Larval @ http://www.textuality.com/Larval (validating parser)
• SP @ http://www.jclark.com (validating SGML parser that does XML as well)
• Expat @ http://www.jclark.com (C parser used in PERL's XML support, and Mozilla)
• XML4J @ http://www.ibm.com/xml (validating parser)
• List of parsers @ http://www.xmlsoftware.com/parsers

• Network bandwidth - in the short term will use a lot more as we must now pass the meta-data around (DTD's or DDML's).
• Processing - think of it, dozens of applications/servers exchanging data with XML and that's just your servers...
• Storage - this one may be neutral, as XML will draw from current databases

XML is definitely cool. If we are ready with the bandwidth, it will be a great millenium