On 2/19/2012 8:32 PM, Fred Cisin wrote:
On Sun, 19 Feb 2012, Jim Brain wrote:
At least in contemporary IT environments, a
mainframe is defined as a
machine that will natively execute 370 assembler. Thus, Hitachi and IBM
manufacture mainframes. I know that's not a good definition, but it's
pretty common.
It's a good EXAMPLE, but . . .
1) It is NOT a definition
2) There exist mainframes that do NOT execute 370 assembler.
(IBM shops may deny that, but it IS true.)
3) There exist NON-mainframes that DO execute 370 assembler.
(add-on boards for 5160, do NOT change an XT into a mainframe.)
I think I can speak
definitively for contemporary IT environments, which
is how I started my sentence. I did constrain my statement
appropriately, and I did note that the definition was not good, but it
is the truth. I have no issues with folks being pedantic, but the above
just smacks of knowledge "showboating".
And, yes, for those IT shops that remember the "mainframe on a card",
they would indeed call that a mainframe. MIPS are not important to the
definition. The solution also need not run zOS, as many people run all
zIIP or zAAP modules and they still call it a mainframe.
As for the "ClearPath" response, I can only say that in all my years at
IT functions, I have never seen a product or service referring to such a
machine as a mainframe. I think in contemporary IT, it would be denoted
as a "ClearPath Server", so as to avoid confusing folks who do indeed
expect "mainframe" to be a 370-based machine (unless the ClearPath runs
370 assembler as well, but with it's Burroughs heritage, I would assume
not).
Like I said, it's not a good definition, but it is the one that is
commonly used in IT today. If your solution will natively execute 370
assembler, people would refer to it as a mainframe. Actually, the IT
advertising folks would call it "legacy", or "big iron", depending on
the advertising context.
Jim