[Info-vax] Whither VMS?

[Roy Brown]

In message <IpKdnQb4e8qwFE7XnZ2dnUVZ_qudnZ2d at giganews.com>, Richard B. 
Gilbert <rgilbert88 at comcast.net> writing at 15:51:21 in his/her local 
time opines:-
>Roy Brown wrote:
<snip>

>>  You could argue that the VAX is superior to the HP3000 in several 
>>ways -  people don't by and large do graphics on the HP3000 like they 
>>do on the  VAX, DBMS has a few features I'd like to have seen in MPE's 
>>TurboImage,  and I'm told that clustering lets VAXes scale more 
>>nicely. But perhaps  VMS file structures are needlessly(?) more 
>>complex than MPE ones -  though both are more so than UNIX bytestreams.

>Hardly anyone does anything on the VAX any longer.  The Alpha 
>architecture pretty much replaced the VAX.  Itanic has either replaced 
>the Alpha or will replace it except in our hearts! <sob>

You reveal my technical limitations as well as my verbal imprecision :-)

I should have been more careful to say 'VMS' in a few places where I 
said 'VAX'. I know the VAX, as a hardware platform for VMS, is pretty 
much laid to rest.

But I've never been a hardware guy; hardly even a software guy beyond 
the point where I opened and used the Intrinsics manual.

So I guess while the HP3000 was going through about 6 hardware 
reincarnations, and those who were close to the metal were having to 
relearn a whole bunch of stuff, I was happily working to a largely 
unchanging API, except for the enhancements, and for when HP switched 
from what it called the Classic architecture (specific to the HP3000) to 
the Precision architecture (PA-RISC, and shared with the HP-UX boxen).

That, I guess, would have been like the VAX/Alpha switch - a 
comprehensive Compatibility Mode where you could run all your existing 
code, binaries unchanged, plus a new Native Mode with the promise 
(fulfilled) that you could recompile your source code with minimal 
changes to run most efficiently in the new environment.

And then, perhaps, with more extensive changes to really exploit the new 
environment - like using a 4Mb array in a COBOL program that used to be 
limited to a 32k stack, and eliminating shedloads of I/O to what used to 
have to be workfiles. (Somewhat to the consternation of the HP COBOL 
compiler guys, that 4Mb array managed to stash itself in the executable, 
expanding it somewhat from a few tens of k's, but we found a way round 
that).

But the main thing was, you could go at your own pace on this; and for 
the users, seeing a dedicated green screen replaced by a PC running a 
terminal emulator was the biggest change they ever noticed; and that 
change was totally decoupled from the HP3000 changes anyway.

<snip>

>VAX is long over and gone!  It was a "CISC" machine rather than the now 
>trendy RISC architecture.  It just couldn't keep up!

Heh-heh. But those pesky x86s are still CISC; and the Itanic is EPIC (as 
in EPIC fail?) rather than RISC. RISC powers my phone nicely, though.

>> My job here is to help give the remaining VAXes the last rites; find 
>>those applications that are still in use on the 'ageing hardware', get 
>>them ported over or replaced. I point out that they could go out 
>>tomorrow and buy a brand new set of hardware to run the current 
>>OpenVMS  8.3, all available and currently supported; I get strange looks.

>Welcome to the club!

We had a joke on the HP3000:

Q. How many MPE people does it take to change a light bulb?
A. Three. One to change the bulb, and two to persuade management not to 
change the whole damned light fitting...

-- 
Roy Brown        'Have nothing in your houses that you do not know to be
Kelmscott Ltd     useful, or believe to be beautiful'  William Morris