🚀 go-pugleaf

In article <809039446.1190@ksmith.com>, Keith Smith <keith@ksmith.com> wrote:
>In article <davis-2008951945510001@pppsb16.itg.ti.com>,
>Tom L. Davis <davis@adrs1.dseg.ti.com> wrote:
>>and true multi-tasking years ago. Big mistake, but maybe not as bad as
>>IBM's 640K error.
>
>Bill Gates 640K error.  All that 640K limitation in MS-DOS was due to
>software not hardware.  Bill admitted he made a mistake when he said no
>one would ever need 640K of memory in a PC.

Are you sure about that? THe 640K limit stems from the fact that AT-bus
devices have memory-mapped I/O in the upper 360K, as well as ROMs, etc.
Wasn't it IBM who put the video memory of the Monochrome Display Adapter
of the PC at location 0xB0000?  That's at the 720K barrier.  EGA and VGA
frame buffers start at 0xA0000, the 640K barrier.

I may have forgotten much of the PC history (good riddance), but I don't
see how Gates could have influenced the placement of the hardware memory
apertures.  Did he have some influence in the matter?  I'd definitely
point the finger at IBM in this case.  Gates can't be blamed for
_everything_, now!

In any case, the 640K "barrier" isn't a problem anymore, since operating
systems can simply map virtual memory pages around the I/O areas. Everyone
in Linuxland sees a flat address space, with the kernel mapped somewhere
in the the upper GB, protected from user processes by a segment limit.


-- 
Finger c2a192@keats.ugrad.cs.ubc.ca for PGP public key. Or contact
one of the public key servers -- the key id is 0xD3C7995D.