> ...
> > After all, just because you don't sell a lot of something profitable
doesn't
> > mean it's not profitable to sell it.  For example, GM doesn't sell a
lot of
> > Corvettes compared to, say, Cavaliers, but they still sell the
'vettes.
>
> But GM won't try to talk you into buying a Cavalier instead of a
'vette :(
 
It will be interesting to see what happens to car models when cars
require operating system software. It may just become too expensive to
offer so many models. If sufficient complexity existed in the GM car O.
S. to permit customization by model within a single code base for the
O.S, then no doubt some other car company's O.S. dedicated to a
particular model would outperform GM's.
 
HP appears to believe that HPUX is the "right" choice for nearly
everyone and that they can meet nearly anyone's needs for specialized
behavior by offering applications on top of (or integrated by
specialized interfaces) the O.S.? This ignores the costs of the
resulting complexity and the level of technical expertise as well as the
time and effort required for support staff.
 
Another thought, HP seems to think redundent headware systems is the
answer to reliability problems, but how can this address software
reliability issues? With increasing complexity, software becomes harder
to use effectively, as well as more expensive to maintain and support.
 
The open systems and PC software trends seem to point toward general
purpose operating software that does nearly everything more or less
acceptably, though perhaps not well. You can get specialized software to
work better in a specific domain, but with this model there is no
alternative for those who need "better" general purpose operating
systems.