On Fri, Nov 16, 2012 at 9:11 AM, Michael Kay <mike@xxxxxxxxxxxx> wrote:
> The second part of the answer (which I guess Ken assumed you already knew)
> is that having started your transformation without an XML document as input,
> you can then read the text file using the unparsed-text() function.
>

Oh yes and he anticipated my laziness in planning to feed the
stylesheet as the dummy input doc. as well.

I've been reading
http://www.saxonica.com/papers/ideadb-1.1/mhk-paper.xml and wondering
whether  the signature  for analyze-string is lifted from the Little
Schemer (Ch 10).