Tony said:
"While we have so many intangibles in the equation, proponents
of either methodology will always be able to 'prove' they are
right and we are wrong or vice versa, so why don't we look at the
grey fuzzy areas and try for a bit of differentiation."
It's worse than just the "intangibles" ... the costs matrix differs
for every shop. The picture for a Windows shop with mostly
'new' systems is quite different from that of a mixed-platform
shop with mostly older systems. A Windows shop generally can't
go to Sun, Linux, or Mac without a lot of training (or new-hire)
expense while a mixed-platform shop can shift their make-up
with much less training cost. Likewise, a shop with mostly newer
Windows servers gets no hardware discounts by switching while
those with older servers can extend the life of that hardware
essentially for no cost.

Each situation is different. In some cases one solution is obvious
and sometimes it's close enough that 'personal opinion'
becomes as important a deciding factor. You can't REALLY make
a decision based entirely on TCO, ROI and the company's "five
year plan" are possibly more important. TCO should be no more
than a guide.

"We are right by Microsoft or we are right by Open Source."

If you are being honest, you will be 'right by the facts'.
Sometimes Windows IS the best choice, based on costs.