On Mon, 20 May 1996 21:07:13 CDT Richard Gambrell said:
>These features, together with the reliability of the machine *and* O.S., fast
>database technology, binary compatibility for 15 years (correct count?)
 
At least 20 years, with a few minor exceptions (APL, COBOL68 runtime, AGL)
 
>Both "problems" can be fixed by marketing, "belief" in the platform by
>HP, and enough R&D dollars to keep the 3000 within the mainstream of
>communications, networking, software development, and processor technology.
>This is much easier (and cheaper) than fixing the 9000.
 
As Ken (I think) pointed out, the unix market may have peaked, or at least
no longer the holy grail.  NT is gaining as people seek escape from the
confusing unix interfaces and the "reality" of running the beast; NT has no
shell, has a GUI, and has Microsoft's market clout.
 
I was going to save this for another target, but what the heck, it still
lacks the punch of a home run hit that I wanted so I'll throw an attempted
ground-rule double at you (speaking of analogies, where's Alfredo hiding?)...
 
The 3000 sort of reminds me of NASA's Saturn V booster that took us to the
moon.  Took us to the moon and back multiple times.  Even Apollo 13 got back.
Sure there was the Apollo 1 disaster, but look at the 3000/CX :-)  I can't
watch one of those documentaries with decent footage of a liftoff without
getting a lump in my throat.  Seven and a half million pounds of thrust.
Liftoff!  Go kick some lunar butt!  Then came the shuttle program, not that
it doesn't have it's fine points.
 
The Saturn V was forgotten.  The designs and the tooling have been lost in
in inexcusable bout of stupidity.  We can't build another one.  NASA needed
a heavy-lift launcher after the shuttle proved not to be *that* economical,
but we sat around while the Russians launched Mir aboard their new heavy-
lift launcher (name escapes me, sorry).
 
So much technology (well, admittedly dated) lost.  Some applications are so
critical there may only be one solution to go that extra step and reach your
goal.  But once that goal is accomplished, don't let it go.
 
I don't think it's a big secret I'm an avid scuba diver, and yes, we
have dive computers.  I use one, but don't necessarily trust it; that's my
life at worst or my health at best relying on those cryptic numbers showing
elapsed time, max depth, current depth, decompression ceiling, and other
data that fits our best-guess differential model of nitrogen absorption in
the human body.  There are lots of dive computers, with varying bells and
whistles and features, and varying degrees of conservatism in their own
decompression models.  Do you want the low bid?  Do you want the least
conservative to maximize your supposedly safe dive time?  Do you want the
most conservative to maximize your safety but decrease your dive time?  Do
you want an air-integrated one to monitor your tank pressure too?  Do you
want a "hoseless" air-integrated one with transmitter/receiver?  The more
gimmicks you add the more can go wrong if, heaven forbid, your batteries
die, or your computer floods out.  No thank you, I still use old-fashioned
tables as a backup and sanity check.
 
If I had a choice of a ride on the shuttle or a Saturn V, I want the latter.
If a dive computer were Unix-based or MPE-based, I want the latter.
Not sure of exactly the moral of this story <grin> unless it is looking
beyond the immediate application need into the foundation of your choice.
 
Jeff Kell <[log in to unmask]>