[Info-vax] OT: news from the trenches (re: Solaris)

[lists at openmailbox.org]

On Fri, 13 Mar 2015 07:46:39 -0700 (PDT)
johnwallace4--- via Info-vax <info-vax at rbnsn.com> wrote:

> OK, thanks for that. Rather terse answer here as I need to be elsewhere...

No fair! Ask and run!

> 1) I do see where you're coming from. 
> 
> 2) Much x86 carp is, as you acknowledge, hidden from most users, admins,
> programmers. But not all. Depends on hardware and depends on OS, and
> depends on your area of interest.

Yes, that's it. I don't believe anything I wrote is important to OS
customers generally. We're talking about the shortcomings of Intel and there
are plenty. But I do feel that at least in my professional circles it is
universally agreed Intel is commodity crapware and I think that part does
affect VMS's future since it is going to be targeted at a platform which is
known to be subprime in the industry.

But the platform does have an effect on the developers that write the OS
and products that run on it. I think Intel brings out the worst in those
groups, for a few reasons. One, it's a sloppy non-architecture so it
promotes nearsightedness and good enough is good enough solutions. It's low
end so people realize it's not worth putting out the best quality anything
since Intel is about speed uberalles and has a sad history of breaking
upward compatability. If they wanted quality they would go with another
platform. That's why Intel seems to me a very bad marriage for VMS. Two, I
think the way things are working today are the tail wagging the dog. If you
think about OS and hardware combinations that developers loved you're going
to think about closely coupled hardware and OS that were designed from the
beginning to work symbiotically. The way things are today that process is
very badly broken in most mainstream OS- basically it's Intel and AMD tying
to outdo each other on speed or other features and then the OS people
usually have to exploit this or that new feature just so they can say they
did. This is rats on a treadmill and not nearly as productive as having the
hardware and OS design in house and having the hardware guys providing what
the OS needs, not the hardware guys driving the OS. We see other successful
ecosystems where the OS and program product needs drive hardware features
and this is the way it should be.

The way things work today with Intel there are several major OS running on
it and the hardware is optimized for all of them then it's not optimized for
any of them. With Linux that's fine because it really exploits little in
the way of hardware on most platforms, with Linux it's more about running
everywhere than running well anywhere. But other OS that *only* run on
Intel like Windows and OS/X are chasing the hardware. That isn't how things
should be done.

Adding VMS to the mix and what do you get? Do you really think what VMS
needs is going to be a priority with Intel? No, we have already heard Intel
has all the great stuff VMS needs. And when it doesn't, then what? What
choices will VMS have when running on commodity hardware where it is a
small fish in a big pond?

VMS needs to run on a premium platform and control the whole ecosystem.
That is the only successful model (I mean in terms of longevity and
quality) we have seen. And that includes Wintel which because of Mutually
Assured Destruction (look at Intel's preliminary financials this quarter-
the warnings are coming) Intel and Microsoft work together. The others, less
and less so.

> 3) You're new to VMS. Welcome. Before drawing conclusions on how much
> of this legacy-x86 baggage will be visible in a nuVMS world, maybe get
> yourself a bit more familiar with how the architectural differences
> have been hidden (or not) in the previous VMS ports? Maybe? VMS is not
> UNIX. Compatibility has been a long term goal (though sometimes long
> term compatibility brings challenges too; a change of platform could
> be a good time to look at some of those things).

Thanks and again I agree the OS end-user or application developer is not
affected directly by these issues although I do believe there is fallout
both tangible and intangible as I mentioned, and it's significant.

Anyway, for somebody who writes systems code in assembler what interested
me in VMS was that it ran on platforms I have no knowledge of, and that are
also supposed to be good and now our hopes are dashed on the rocks on Intel
crapware again. It's depressing to see what a small world it's becoming,
where everybody talks about choice but there really is almost none. Anything
we can do to change that is A Good Thing.

> And then once you know how much platform-specific stuff has historically
> been (in)visible, we can start guessing how much it will *matter* to the
> people with the chequebooks.

Agreed.

But I'll continue to run old versions of VMS if I do at all, because I have
no interest in spending any time with Intel.

-- 
Please DO NOT COPY ME on mailing list replies. I read the mailing list.
RSA 4096 fingerprint 7940 3F02 16D3 AFEE F2F8  ACAA 557C 4B36 98E4 4D49