VMS versus UNIX

[Daniel Fernandez Illan]

This is a very interesting link about VMS comparisions:
http://h71000.www7.hp.com/openvms/whitepapers/alinean_tco.pdf
LONG LIFE TO VMS!!!
Saludos,
Daniel

[Robert Gezelter]

Daniel,

Thank you for bringing that whitepaper to our attention. Personally, I had not noticed that paper on TCO, although I have cited some of the other TCO papers extensively.

- Bob Gezelter, http://www.rlgsc.com

[Ian Miller.]

It's another good one. It would be interesting to see if more people would read it if it did not mention VMS until the end. Some people are still stuck with the impression that VMS is a dead O.S. and therefore will not read articles with VMS in the title :-)

There will always be people irrationally biased against VMS or biased towards Microsoft or other, but more whitepapers like this can help to convince the sane ones.

[Jan van den Ende]

Yeah..

Although I wonder WHO did the testing of running VMS 8.2 on Integrity for five years??

Was this somehow "tranlated" from earlier versions on AXP, or was it extrapolated from 5 days (the report was from januay, and 8.2 was only released mid-january).

In my view, it would have been more convincing if they had used real-world measurements...


FWIW

Proost.

Have one on me.

jpe

[Wim Van den Wyngaert]

Not impressed.

I'm sure that you can find documents on the IBM or Sun site that prove that their platform is cheaper.

Take the "change costs" on page 17. Average application porting cost : 0$ for VMS. Then they should come and look to our applications ... HP estimated 100 days and it became 1000 days (env and applic).

Wim

[Willem Grooters]

Nor am I - for Jan's reasons.
nevertheless, it holds some truth.
My two cents:

* I once did the techical maintenance of an application (based on an (at the time) HP product, using Oracle 6) on a smaller HP machine running HP-UX9. When upgrading to the then latest PA-RISK platform, this required the lastes OS version, the latest version of the product and a next Oracle version. It has taken me a year to prepare and test, and one week to get the whole thing running smoothly. (the transfer of data took the whole weekend, but the backlog of processing required a week to die out)

* SUN (and IBM) keep telling that conversion to their platform is cheaper. What they do not count is the cost to port the application, cost of data conversion, cost of inavailability, and the risks involved.
Any good OBJECTIVE whitepaper to be able to calculate these is welcomed.

Willem

[Ian Miller.]

find yourself a copy of the Unix Haters Handbook originally published in 1994 by IDG.
Its available as a pdf file at various places.
e.g
http://web.mit.edu/~simsong/www/ugh.pdf

http://catalog.com/hopkins/unix-haters/handbook.html

This manual explains exactly why unix is a really bad idea. Yes its dated but is it wrong?

[Ian Miller.]

and here is a related rant
http://www.mindspring.com/~blackhart/I_Miss_VMS.html

[Daniel Fernandez Illan]

Then, a reflection on this threat.
If like it seems to show linked document, VMS has more stability, more safety and major facility of control.
Why does proportion grow between the systems UNIX installed and the VMS for big projects?
Might it be a problem of formation of new managers?
Lack of information about operating systems?
Minor develop of freeware software to VMS?

Saludos.
Daniel.

[Wim Van den Wyngaert]

Unix : Sun, HP, IBM, ...
VMS : HP

All proprietary OS (MPE, AS400, MVS) are being banned ...

Also : Unix sources can be exchanged over Unix platforms without big changes. When they visit VMS some rewriting must be done to make it work/performant. Also : you need to find someone for doing (and keep doing) the job.

Also : DEC did bad marketing 15 years ago.

Wim

[Jan van den Ende]

Wim,


Also : DEC did bad marketing 15 years ago.


Make that: DEC (and follow-ons) dir that for the _DURATION_ of (at least) 15 years,
and I fully agree!

Proost.

Have one on me.

jpe

[Wim Van den Wyngaert]

Also : Unix is a lot faster (cache, dirty IO's, RMS slowdown, file extention slowdown, working set extention slowdown, ...).

Wim

[Daniel Fernandez Illan]

Win

Unix is a lot faster

25 years ago I worked with a OS specific to Real Time applications (MPX-32). Yes it was a very fast OS, but controls over operations were very bad, if you defined in FORTRAN 77 an array IEX(3) and index had a non valid value (index=-32767) when you intended assign value of array IEX(-32767) then system crashed.
Yes, very faster but very bad mistakes or errors control.
Saludos.
Daniel

[Ian Miller.]

"Unix sources can be exchanged over Unix platforms without big changes."
Only if carefully written otherwise porting from one version of unix to another version of unix can be a lot of work.

The mistaken attitude is that unix is not propritary. Most unixes belong to somebody and they are all different but have the same flaws.
You get a system running on one unix flavour and theoretically you can move it but in practice its a lot of work.

I guess the answer is market perception - where are VMS Marketing? VMS is complient with more standards than a lot of unixes (posix etc) but nobody notices.

[Jan van den Ende]

Ian.


.. but have the same flaws ..


If THAT were true, things would be easy in *UX land!
No, they all have DIFFERENT flaws (only mostly the same TYPE of flaws), so you have to adapt separately to each different one.

Proost.

Have one on me.

jpe

[Willem Grooters]

Wim,


Add these causes:
No checks on privilege, protection, security, buffer/stack/data area overflow, invalid concurrency (locks), OS/system health oin evry level. No real error control.

(well, some may have improved but in my yougre days these were certainly true)

Willem