A significant difference between z/OS and Windows environments is the
management philosophy in the z/OS environment that all the core code
that controls the system behavior and integrity can (and should) be
completely isolated from application and third-party code in data sets
that are protected, and that the hardware architecture and SAF product
enforce that code separation. Acceptable maintenance practices only
allow installation of system code at times of the installation's choice
on non-production data sets, and via means (SMP/E) that allow the
affected data sets and the status of every module before and after the
change to be known.
If zOSMF ever becomes such a black box that that SysProgs lose the
clarity of what data sets must be carefully protected as critical to
system integrity, that would be badness.
Windows core system maintenance occurs mostly at the whims of Microsoft
on a running system and there is no way to track what individual files
or directories may be changed or easily undo individual changes. It is
still possible for third-party product installs and maintenance to
update files in directories shared with Windows core code or potentially
modify Windows system files, and they can also update the same registry
file that is essential to Windows operation. At least Windows asks for
permission to proceed with a product installation, but if the product is
one you need, the only choice is to respond "yes" and pray it doesn't
change anything it ought not. This lack of positive protection over all
Windows core code is why a Windows re-install is sometimes the best
repair choice.
The need under the PC architecture to support a wide variety of
vendor-unique hardware devices with hardware vendor device drivers that
run at the Windows kernel level is another major potential integrity
exposure in Windows.
Given the number of integrity exposures that Windows allows out of
Microsoft's direct control, It's amazing Windows is as reliable as it is.
Joel C. Ewing
.On 5/27/23 04:39, Jack Zukt wrote:
I can understand the windows and android analogy. And I do hate not being
able to solve windows and android problems with the ease I solve them on
z/OS.
Over the years colleagues from other areas have come to me for help with
problems on their turf, and I was able to help them because I knew how
things worked between them. Usually I could point them in the right
direction because I understood MVS. I did have to learn a bit of windows, a
few years ago, so I could solve the problems that my father had on his pc.
But sometimes I would have to reinstall windows to solve them. And that is
not an option with z/OS. So, I suppose we are going into a new way of
working that will be fully automated installations and IA maintenence. I
did try zosmf a few years ago. I did not like the underneath complexity
then and I still do not like it now.
This is supposed to be a tool to install anything and everything, and to be
a worflow to multiple maintenence tasks. The amount of RACF definitions and
authorizations are staggering.
Sometimes I do wonder if the way I look at zosmf has more to do with my age
than with the product itself. I delegated the zosmf implementation to my
much younger colleagues, so that my prejudices would not get in the way.
But what I can say at this point is that they have managed to install it
and, I think, they have managed to use it. However they have learned
nothing that will help them to understand what is beneath it. They still do
not grasp basic z/OS concepts. I still think that the best way to learn MVS
is to build one from scratch. Unfortunately that is something seldom done.
But moving your systems to a new server pack is the best alternative.
There is no point bitching about zosmf and the good old days, as this is
the way IBM is building it now. But Windows and android are not stable and
reliable systems. At least, not the ones I have running on my devices.
Jack
On Fri, May 26, 2023, 17:38 Phil Smith III <li...@akphs.com> wrote:
Jack Zukt wrote, in part:
The real problem, as I see it, is that drag and drop interfaces move you
away from the need to know what you are doing.
That’s the *goal*. Do you know what the Windows installer is doing?
Android? iOS? No you do not, beyond the high-level “putting **** in places”
and “making it bootable”. While it makes me nervous because, like you, I’ve
been doing this for too long, most people see this as a good thing, and I
can’t really dispute it.
In 1920, if you had a car, you understood air/fuel mixture, how to do oil
and tire changes, etc. Most people don’t now, and that’s not a bad thing.
Sure, occasionally they get stranded, but by and large, it Just Works.
On the flip side, I was switching phones and the migration failed. I did a
reset of the new phone and restarted it, and it worked the second time. Did
I understand the process? No. Did I need to? No. Am I happy that I couldn’t
tell what failed? Not really, but, again, I really didn’t need to. And
99.44% of people wouldn’t care, as long as it worked the second try.
and Colin Paice asked:
Would it help if we moved to standard configurations?
That’s the point I made before: with Windows, you’re forced into a fairly
standard configuration. With z/OS, it’s a bit late (by almost six decades):
sites aren’t going to rearrange everything. For new installs (all handful
of them), sure. But that doesn’t really help at this point, alas.
...
--
Joel C. Ewing
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
