On Mar 25, 2005, at 8:58 AM, Dave Tweed wrote:

> I'd say that Microsoft has set the PC industry back a good 20 years
> from where it would otherwise be -- look how long it took them to
> decide that virtual memory and preemptive multitasking were Good
> Things. These were known in the 1970s (on mainframes and 
> minicomputers),

It wasn't just microsoft, of course.  The microcomputer industry
in general completely ignored a lot of stuff that had been learned
on mainframes much earlier.  Arguably, microcomputers set back the
state of computing 20+ years, and it wasn't microsoft...  (of course,
it would have been nice if microsoft, as the industry leader, was
more aggressive about implementing known technology.)

OTOH, we all tend to neglect the things that microsoft has done well.
The degree to which windows will install and run on a HUGE variety
of hardware without expert hacking is amazing.  The amount of backward
compatibility is amazing.  Their support of hardware vendors is amazing.


> microprocessors had the hardware to implement them in the early 1980s, 
> but we didn't get them in Windows until the mid-to-late 1990s.

Early attempts at virtual memory in microprocessors were pretty
bad.  For some reason, hardware vendors were implementing segments
even though most 'real' OSes would have rather had paging.  The only
paging micros in the early 80s I can think of were the early SUNs,
and that was all in external logic and memory.  The 80386 and 68030
were about the first micros with decent MMUs, and those were pretty
much LATE 1980s.  (and don't forget the famous bug in the 68000 that
had would-be unix workstation vendors doing thing like using a extra
CPU to handle page faults; and that was just interfacing to EXTERNAL
MMUs.)
-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist