M$ adds on Linux Today

[Kurt Wall]

On Sun, Oct 16, 2005 at 06:11:40PM -0700, Bill Campbell took 52 lines to write:
> On Sun, Oct 16, 2005, Collins Richey wrote:
> >
> >Yeah, MCP was a dream of an operating system - almost no maintenance -
> >it just worked. The only problem with the Burroughs medium system
> >boxes was the fact that they were decimal arithmetic only. If you
> >needed binary for anything (CRC, etc.) you had to write binary
> >simulation code!
> 
> That was true for the Medium Systems (B-2500->B-4800), but not
> for the Large Systems (B-5500->B-6800) and A Series machines
> which were at least 48bit binary words.  The B-1700 series were
> more interesting in that they have a variable microcode that
> switched depending on the language, decimal for COBOL accounting,
> binary for FORTRAN, etc.  I never had a chance to work on the
> B-1700s, but thought they would be very interesting.

Those A-Series machines were absolute beasts. I thought of them as
monstrously huge backplanes with a bunch of CPUs attached. Mainframe
had parallel processing and host partitioning a *long* time ago.

> I cut my mainframe teeth on the B-5500s, programming mostly in
> Burroughs extended ALGOL, which was the system language -- there
> was no Assembly.  The B-5xxx series were the first to have real
> virtual memory, almost a decade before IBM invented thrashing and
> called it VM.

Oy. My second programming language was WFL, WorkFlow Language, that 
stripped-down ALGOL derivative that Burroughs (and Unisys after them)
used for job control. As a result of WFL, I had no trouble learning
Pascal.

Kurt
-- 
Peter's Law of Substitution:
	Look after the molehills, and the mountains will look after
themselves.