An important question may be whether you are doing the XSLT transform
in .NET code, or using a stand-alone XSLT program such as msxsl.exe.
If it's the latter, then Brian is correct: your input must be properly
formed, which means that old-style HTML that doesn't conform to XHTML
won't work, because the tags aren't matched. For example, most people
writing HTML don't bother putting in the closing </p> tag at the ends
of paragraphs. They just put <p> all over the place and expect that the
browser will realize that <p> in the middle of another <p> really means
</p><p>: close the previous paragraph and start a new one.
As well, lots of people writing HTML don't sweat about whether to use
<p> or <P>: they're both the same thing in HTML, but not in XML.
So, if you're using a stand-alone engine then you have to feed it
well-formed HTML, which the industry calls XHTML.
If you're reading the HTML into .NET and then transforming it within C#
or VB, there may be a way, as Martin pointed out, to massage the
resulting data structure into something the XSLT transform engine will
accept.
