This seems to be exactly as Micheal Kay described it.  Hope that helps...
Benjamin

From: Brook Ellingwood <brook@xxxxxxxxxxx>
Reply-To: xsl-list@xxxxxxxxxxxxxxxxxxxxxx
To: <xsl-list@xxxxxxxxxxxxxxxxxxxxxx>
Subject: Re:  xsl Greater than and Less than.
Date: Mon, 06 Oct 2003 10:29:52 -0700

If I can interject here, exactly how strings are compared in XSL 1.0 has
been an area of some confusion for me. Doug Tidwell's O'Reilly book states:

"If both objects are strings: Then they are equal if their Unicode
characters are identical. For less-than and greater-than comparisons, the
character codes are compared."

I've been told by people at Microsoft that this interpretation of the spec
is wrong but that early versions of MSXML (not sure which ones) used this
same mistaken interpretation, but MSXML4 doesn't. At the time, I tried to
figure it out what the spec was saying for myself, but didn't get very far
-- I was a bit out of my element at that point. It sounds like this agrees
with what Michael is saying.

-- Brook

> From: "Michael Kay" <mhk@xxxxxxxxx>
> Reply-To: xsl-list@xxxxxxxxxxxxxxxxxxxxxx
> Date: Mon, 6 Oct 2003 16:46:31 +0100
> To: <xsl-list@xxxxxxxxxxxxxxxxxxxxxx>
> Subject: RE:  xsl Greater than and Less than.
>
> If I decode your message correctly, you are using whatever processor
> Microsoft Windows gives you. That's an XSLT 1.0 processor. With a 1.0
> processor, when you compare two strings using < or > they are converted
> to numbers, which in this case gives you NaN, and NaN<NaN is always
> false.

XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list