[Info-vax] Gartner report on VMS future.

[ChaosLess]

On Sep 20, 2:31 am, Michael Kraemer <M.Krae... at gsi.de> wrote:
> ChaosLess schrieb:
>
>
>
> > These libraries today achieve 50k lock
> > ops/second on 4cpu Opteron Quad (Solaris10), are fully multithreaded,
> > provided scalable AST delivery,
>
> it's probably your business secret how to achieve that ...

no secrets other than tight programming, lots of hard work writing
code that we keep improving.  and reworking parts of the libraries as
technology improves on unix's for it...  how we did async programming
in 1990 didn't work well for async qio and ast delivery was a
challenge, just like stack unwinding...  reworking core components,
adding scalable threading, and LOTS of testsuite programs :-)

>
> > in about 1990, we had the same market 'guru' types telling us we were
> > in a doomed market
>
> Why? Every customer jumping off VMS would have been a potential
> customer for S7. Given the number of VMS users was
> larger by orders of magnitude back then, this would have been
> a lot of business.

so we did hope, and yes a lot of companies have come through our
virtual doors.  we found that helping them get hands around the
decision process was at least, if not more, important even than the
software we wrote/write to (try and) make it possible.

>
> > that would end with the demise of VMS in the very
> > near future.
>
> What's wrong with that prediction if you compare
> the wide spread use of VMS in 1990 with its
> state a decade later, let alone now?

those predictions were the fat lady would be singing in 'a few
years.'    imagine trying to predict longevity of a business with that
backdrop...   the truth, as many here know, is that the rush to open
systems was a move towards raising the commodity level of hardware and
o/s up another level.   while that did decrease the price point in
many respects, this was/is often a confusing distinction when the
purchase is about perception and often not made by the most technical
of criteria or individuals.   VMS shops we encounter now are often the
very definition of 'mission critical' and in hindsight makes me glad
we've had 20 years to (try to) prepare for their needs.

we've found many companies still on VMS rely heavily on some very
strong features that do not easily migrate to other platforms.  we
cover most all system services, library and runtime api calls, dcl,
jsp, etc... always with ever increasing fidelity.   our offering those
as software-only sales is relatively new and a result of our (attempts
at) maturing the code base we started more than 20 years ago.

some areas we still only have barely touched, like true clustering
support, and we are always trying to weigh up the difficulties in
their implementation especially in anticipation ahead of the next
customer's needs.  i know our relationship with this group is
sometimes turbulent but we really are working towards letting VMS and
*nix systems coexist on VMS terms.  we see that as extending the life
of VMS design paradigms and programming features, and of course
customer software, that are still unique and valuable.