XML's many applications include one traditionally handled by HTML,

document source. Sometimes you want to include some HTML in your XML

source; particularly when styling your XML with XSL to HTML, you may

want some literal HTML in the output. In some cases, you'd just like to

include some well-formed mixed-content HTML in your XML, and not worry

too much about controlling its structure.

Writing explicit XSL rules for each HTML tag would be tedious and hard

to maintain. This week, I'll present a quick shortcut that allows you

to include arbitrary mixed-content HTML in your XSL stylesheet's output.

The XSL stylesheet's input includes some XML document, constrained by

(that is, validated against) a particular DTD. To define the element

containing the arbitrary HTML's content model, simply define content

models for each of the HTML elements you want to use and their

attribute lists. For example, define the

this:

?!ELEMENT a (#PCDATA)*?BR>

?!ATTLIST a

href CDATA #IMPLIED

target CDATA #IMPLIED

name CDATA #IMPLIED?BR>

(This example is simplified, since it allows only unformatted text

inside the hyperlink text node.) Create a definition like this for each

HTML tag you wish to use. After you've defined the tags, define a

parameter entity that includes all of the HTML "pass-through" tags

you've defined:

?!ENTITY % htmlpassthru

"a|i|b|code|br|tr|td|th|img|font|em"?BR>

Anywhere in the DTD you want to include mixed-content HTML, use this

parameter entity in the DTD. For example:

?!ELEMENT HtmlHelpText (#PCDATA|%htmlpassthru;)*?BR>

Now for the stylesheet. Defining a separate rule for each possible

HTML element within a HtmlHelpText element would be extremely tedious.

Fortunately, a single rule can handle all such elements. Define an XSL

rule that matches the tag name AND uses

entire node (including attributes) to the output.

?!-- EXAMPLE XSL stylesheet --?BR>

?xsl:template

match="a|i|b|code|br|tr|td|th|img|font|em"?BR> ?xsl:copy-of select="."/?BR> ?/xsl:template?BR>

?!-- End EXAMPLE --?BR>

Any nodes in the XML input tree with these tags will be copied to the

output structure unmodified, attributes and all.

This technique is somewhat "quick-and-dirty". Choosing XHTML, which

reformulates HTML as XML and brings XML's extensibility and structure

control into the world of HTML, would be a better solution. But that's

a topic for another newsletter.