On Thu, 9 May 1996 07:07:30 -0700 Bruce Toback said:
>Jeff Kell writes:
>> Even IBM's
>> VM/ESA which does emulate multiple virtual physical machines (albeit all
>> IBM platforms) only allows them to co-exist, not co-operate.  They made a
>> really big deal not that long ago about "shared code segments" that multiple
>> users could share, and elaborated on converting your code to localize all of
>> your data references to your assigned "data area" (non-trivial).
 
>VM is in a quite different class from MPE's code and data management. In MPE,
>the code and data protection can be overridden by privileged programs. You
>can't run an OS as a process under MPE without compromising system integrity.
>The point of VM is that there must be no way that a program can tell that it's
>being run in a virtual machine, even if the "program" is a privileged OS.
 
Actually I wasn't referring to shared OS segments, though the points are very
valid.  VM is older than MPE :-) so I was really making references to my past
dues-paying (blues-playing) days as a big blue systems programmer, a career I
no longer care to pursue.  Simple separation of code and data was quite an
effort.  "Re-entrant" (shareable) and "relocatable" (dynamic loading) code
were the result of careful planning and tedious details back in the dark ages.
Then along comes MPE with inherently re-entrant, relocatable, protected code
and even has the gall to use programs and files as swap area!  Geez, even
a COBOL programmer can write shareable code now; such heresy! (OK, that's
my quota of exclamation marks...)
 
Jeff Kell <[log in to unmask]>