I can't read your email properly, as I don't use IE. Do I presume you
in effect are describing a file done as follows:

--- "Excel.idx" ---

Title: Many interesting figures which you should read
Description: Includes a solution to life, the universe and everything
Src: excelfiles/interesting.xls

Title: Other information you may not know
Description: Various bits of trivial data by Michael Caine
Src: excelfiles/notalotofpeopleknowthat.xls

etc, etc.

--- end of "Excel.idx" ---

If so, then you could perhaps create a file in the collect/????/etc
folder to cope with it (as one does for other files normally), and
then have a plugin which bluffs some content, possibly even based
on the description or something? I usually take this approach when
building collections based on postscript files, etc., anyway, just
to make sure that I don't rely on uncertain behaviour of any "smart"
code.

The Quark stuff's more involved, at least if you want to index the
content.

I'm CC'ing this to the greenstone mailing list in case it is of help
to others...

--
George Buchanan
Research Fellow, Digital Libraries
Middlesex University, London, UK