Re:
> 1) With the new "A" and "N" class system announcements they have a good
> chance to upgrade many of the old 9x7 M/C's. And looking at the hardware
I'd like to propose three points, which I'll call "run it", "trust it", and
"evolve it".
1) "run it"
The best thing we could do for the future of the HP 3000 is
rally the users to complain about the software crippled CPU of the
A-Class HP e3000. (The 440 MHz CPU is software reduced to 110 CPU)
(I.e., let the users *run* the A-Class!)
HP *says* they want to compete in the internet area ... but the
one thing we need there is a fast CPU. There isn't a single
program that wouldn't benefit from a faster clock on the A-Class.
This includes software that's CPU bound, disk bound, *or* network bound.
The sole reason for the clock crippling is to preserve a CPU-speed
market distinction between the A-Class and the N-Class.
Well, guess what, HP didn't need that distinction on the HP 9000 side!
They let I/O expandability and memory expandability be the
market distinction!
We're not talking a minor difference here, like with the 918 vs. the 928.
We're talking a factor of ** four ** in performance.
I think HP should immediately:
1) announce a 8-user license A-Class with 440 MHz CPU at the $15K level;
2) announce an unlimited user license A-Class with 440 MHz CPU
at a slightly higher price ... say, $20K ?
By limiting the number of users at the entry level price, and increasing
the price for the unlimited number of users, they preserve some of
that apparently-needed market distinction between the A-Class and
the N-Class.
Today, however, HP's message is crystal clear:
We don't want the HP 3000 to be able to compete with *anything* else.
HP can change this ... easily.
2) "trust it"
The second thing the users could do is encourage HP to work on the word "trust".
As in, we trust HP to protect our investment. This would be done by announcing
support for MPE/iX 7.0 on the HP 3000/9x7 computers.
The decision to prevent 7.0 from running on the 9x7 appears to have been
strictly a marketing decision. That's just not right. A lot of users have
9x7s along with newer HP 3000s ... and they're paying for software support.
Now, they'll have to junk them, or start running with a mix of old and
new operating systems.
HP apparently feels that they'll get more money in sales of newer systems
than they'll lose from dropped support on the 9x7. Until support is dropped,
however, it will cost them *more*: the Response Center will have to answer
calls about problems long since fixed in 7.0, 7.1, etc.
Today, however, HP's message is crystal clear:
We'll obsolete your hardware when we damn well please.
HP can change this ... easily.
3) "evolve it"
I've been proposing, for a number of years, that HP add the ability to
run HP-UX binaries on MPE/iX. Although this isn't a trivial task, the benefits
are enormous. We've already seen that nearly *every* port of a Unix program
to the HP 3000 runs into problems (some minor, some major, some insurmountable)
(e.g., apache web server, the unlamented Netscape web server, samba, tracert).
Rather than attacking them piece-meal, I'm suggesting solving all of them.
Yes, HP would have to greatly improve the HP 3000 networking code ... but,
hey, that's years overdue anyway! (That's probably the worst part of the
work, BTW.)
By being able to run HP-UX binaries, the HP e3000 would be able to compete
in every bidding contact calling for "Unix" ... something we can't do today.
Vendors would be able to put their products on the HP 3000 ... and then
could be encouraged to start taking advantage of HP 3000 specific things.
This, coupled with a *REASONABLE* clock speed on the A-Class HP e3000s would
mean that there would be little or no reason to every buy an HP 9000 A,
when the HP e3000 A-Class can do everything it can do ... at least as
good or bettter! And, the HP sales force should quickly realize what we've
been telling them for 15 years: an HP 3000 sale is better than an HP 9000 sale!
Apparently, the worlds largest vendor of proprietary operating systems
agrees with me...IBM recently announced software for AS/400 which allows
the user to run AIX programs!
HP has had an excellent record of evolving MPE ... in recent years, we've
had POSIX, many major IMAGE enhancements, Large Files, and memory > 4 GB.
(We've had some user-initiated & HP encouraged major new software program
added, like Apache, samba, gcc.)
However, the *ONLY* major project HP has said they're even considering
at this point is to run on IA-64, and that's not evolutionary (it's an
example of supporting a new hardware platform, albeit a major project).
The HP-UX binaries on MPE is an evolution project.
Today, however, HP's message is crystal clear:
We're not doing much more than bringing new I/O channels (e.g., Fibre),
and supporting new hardware platforms.
HP can change this: it isn't easy, but it has to be done.
There you have it: a true approach to "run it, trust it, evolve it".
Stan
Stan Sieler [log in to unmask]
www.allegro.com/sieler/wanted/index.html www.allegro.com/sieler
|
