On Thu, 28 Nov 2002, Tony Duell wrote:
[Replacing the ROMs on a PC motherboard)
Why do you think I intend to do this to a 5150 or 5160 board? :-)
An appreciation
for how much nicer they are than the later stuff?
On an old PC,
you could put a magic header and a
checksum on a 2716, and basically load an .EXE file
It's actually just an 8088
machine code program with the header and
checksum. No .EXE file header as well.
.COM style code (without the ORG 100h) would be one of the easiest ways to
do it if one wanted to use "modern" development tools (such as MASM 1.0,
LINK, EXE2BIN)
> capability work. on a "real" PC, you
can remove the
> BASIC chips and do this.
On a REAL 5150, you don't need to remove the
BASIC. There is already an
EXTRA ROM socket!
[Unverified CCUL: IBM asked Microsoft how large the ROM BASIC would
be. Microsoft said, "32K". IBM's engineers figured that things might not
go entirely as planned, so they put in one extra ROM socket "in case it
went over the targetted 32K". The Microsoft people were miffed that their
guess of 32k hadn't been taken as being accurate. So, when they were done
(at about 30 or 31 K), they padded it out with incompletely implemented
stuff to bring it up to EXACTLY 32768 (32K) bytes.]
AFAIK the ROM BASIC on an IBM machine doesn't have
the header, etc. It's
just some code at a known adresss that's exectued by the appropriate INT.
Very early 5150s don't implement the BIOS extensions IIRC, but later 5150
BIOSes and all 5160s, etc, do.
But even [particularly] the earliest had the EXTRA
ROM socket.
'Course it's nice to put your add-ons on an ISA card, so that they can be
switched more easily to other machines, including NEWER stuff, like 5160
(XT) and AT.
--
Grumpy Ol' Fred cisin(a)xenosoft.com