On Wednesday, February 24, 1999 04:52, F. Alfredo Rego
[mailto:[log in to unmask]] said:

> The last paragraph says, "Ultimately, users will have to move
> from Windows
> 95 and 98 to Windows 2000" and previous paragraphs mention
> the bright side:
>
> http://www.computerworld.com/home/news.nsf/CWFlash/990219page1
>
> How does this compare to your experiences regarding MPE
> upgrades on the
> HP3000?  My experiences regarding MPE and the HP3000 have
> been excellent.

Back in the Win 3.x to Win 95 days, it was "Ultimately, one would have to
migrate from Win 3.x to WIn 95" - those were the official words from
Microsoft.  There are still Win 3.x boxes running out there, quite happily
as well.

The reason for the migration isn't a secret, and has been this way for some
time:  A Unified Windows.  The whole strategy was to take the desktop OS
(Win 9x) and the Server/High-Powered Workstation OS (NT), and unify them
under the same code base.  It had to be done in steps though, since they
were, at that time, being worked on by two separate groups.

The first inkling of that came when the Windows 95 User Interface appeared
in NT 4.0.  Windows 98 uses more of the NT code base now but still has a
ways to go.  Gates said over the weekend (or last week, I forget), that
there will probably be one more minor release under 98, name unannounced,
before Win 9x can be truly unified under the NT code base.

Now, this should solve a lot of problems.  Already, most of the Win 9x
programmers are over on the NT team, performing the "unification" with
existing NT programmers, and a handful of 9x programmers remaining to squash
bugs.  There will be an OEMSR1 and a SP1 for Win98 coming out in the not too
distant future - but the goal is to have one OS - Windows 2000 (which is
currently being slowly adopted - just check out the ClubWin Discussion
Group, where the name Windows 2000 is used for NT at the present).

Migration is never easy.  I remember migrating from MPE/V to MPE/iX.  Most
of it went nicely from a migration of code standpoint.  Retraining
operations staff was not so nice though -- even though many of the new
commands were still there, new ones were introduced.  A new and more
complicated boot procedure (I still think that mounting a tape, setting the
switches to %3006, then hit Load/Run was easy - for a Coldstart).  Wording
Changes - Warmstart, Coolstart, Coldstart, Reload and Update became Start
Norecovery, Start Recovery, Update, Install...  Vendors try to make it easy
to migrate over, but it can be difficult.

V-to-iX was easy in comparison to most others.  Windows has so many
different hardware configurations to contend with, so many devices, mixing
and matching those devices, and so many software vendors who may or may not
have written well-behaved code (like not straying from the APIs provided and
implementing their own) - this makes the job of migration much more
difficult.  There will always be one application that will come crashing
down to its knees, no matter how much it's beta-tested.

I don't know if I answered it here - but at least some of the information is
out for comparison.  Upgrading MPE from "V" to "iX" was not terribly
difficult.  Upgrading to IA64, if HP keeps to form, should not be terribly
difficult either - but again - we won;t know until we start seeing some
hardware, and some code from the lab.

Cheers,
Joe (off for his second cup of coffee for the day, and in a better mood than
yesterday :)