Colin Barnhorst said:
Why should Vista only be 64bits? Hundreds of millions of computers are
capable of running 32bit Vista but not 64bit. Tens of millions can run it
with Glass. Why should those users be told they cannot upgrade to Vista
just because 64bit computers are a good idea?
What does "retreat to the 32-bit version" mean? They are using 32-bit now
so what's to retreat to? Who are we to tell anybody that the computer they
have now is not acceptible?
Are we to become the architecture police?
From Microsoft's point of view, it is expensive to maintain multiple
versions of the same OS. So certainly their profit goes up if they can
consolidate. Short term they sell less, but within a year people would
welcome a migration to 64 bits as prices on hardware commoditize.
From the customer's point of view, it is expensive to have to support
different versions of the same OS. No one wants the hassle of multiple
versions of drivers and applications and partially supported configurations.
Every release of an OS involves decisions and trade offs that necessarily
exclude some part of the market from participating in the new OS, based on
memory requirements, processor requirements, graphic requirements whatever.
That puts a software manufacturer in the role of being architecture police
whether they want to be or not. Microsoft made a mistake in 1985 of trying
to support Windows 1.0 on 286 processors. They used all your same
arguments about trying to find a mass market, but in that case the marketing
decision led to a low market adoption of the new technology because it was
not able to use features in the higher end 386 architecture effectively.
In this case supporting 32-bit means they will see rapid adoption of Vista,
but at a higher support cost for Microsoft and the customer's both. I
just think if you were to look out over a four year adoption curve,
supporting 64-bit only would not have significantly lowered Microsoft's
sales. They would just have been highly skewed to years two through four
instead of front-loaded. And customer adoption costs would be lower,
because the human costs of having to research compatible drivers and
applications and support complex configurations always greatly exceeds the
cost of the hardware.