Another way to look at the choice may be in a cross-section of IBM
products, and where RS/6000+AIX is put to heavy use vs. Windows. In
some arenas, IBM makes the realistic business choice to implement on
Windows because that's what some customers are conditioned to want -
be it the choice which makes the most sense or not - and in some
areas IBM makes an outright technology choice for *nix on RS/6000 or
xSeries. In its Infoprint Manager product, IBM offers it in both AIX
and Windows flavors - but the Windows version is the less serious
implementation, where it makes sense in more modest environments,
while the AIX version is the workhorse.

While we performance tends to dominant our thinking as technicians,
it is just one factor. Security is another major factor - where
nothing need be said about the history of Windows in that area. This
leads to another major decision area, or perhaps philosophy, which is
open systems. Do you want to commit to a proprietary operating system
environment, or do you want a product of democracy, with published
interfaces and formats? That choice is much like whether you want to
go with LTO versus AIT, where you have a variety of vendors to choose
from in the former choice. As Wanda alluded, in a proprietary
environment, your ability analyze performance issues may be greatly
limited...where one vendor "knows best", and "you don't need to
know". In counterpoint, in a smaller environment with constrained
staffing, no real systems analysis is going to be possible anyway;
but at least with open systems you have the opportunity to adopt and
implement tools at little or no cost.

In my mind, the open systems choice is the most compelling factor. It
give you choices, where you won't be led around like you have a ring
through your nose - which by serendipitous metaphor makes me wonder
what's going go happen to Windows customers with Longhorn.

   Richard Sims