On Thu, 27 May 2021, Chuck Guzis via cctalk wrote:
I'd been cutting floppy controller code since
about 1975; the weirdness
of the IBM PC design really made for some head-scratching.
I did some FORTRAN since 1967, but until the TRS80, I couldn't own a
computer, and hardly ever got a chance to touch them, and was strictly
limited on what I was even allowed to do.
So, I never talked to the FDC until about 1980. (when I found Bill
Barden's paper about talking to the TRS80 FDC)
Sometime in the 5150/5160 days, I published
instructions on how to
modify a PC floppy controller that used the 765 with a WD926 data
separator. It was quite simple--for some unknown reason, both of the
clock rate select lines were hard-wired. Getting FM support was mostly
a matter of lifting one of those lines and running a jumper to the
MFM/FM output on the 765.
But then IBM and clones did the same sort of thing
with the parallel
port adapter. Changing one to bidirectional (and curiously
PS/2-compatible) was mostly a matter of lifting the output enable line
on an IC and tying it an unused bit in the control port.
Sometimes I wonder if IBM did this sort of stuff intentionally.
I suspect that it was intentional, but through ignorance.
"Should we have FM?"
"Should we make the printer port bi-directional?"
To which some IBM manager with little or no technical expertise asked,
"Do we NEED to?"
and, assuming that it would cost, rejected the "UNNECESSARY expenditures"
for the extra capabilities.
I modified a PC parallel port, and played around with it. My goal,
XenoComm Parallel, was NOT to use it with software on both machines. I had
run into a few machines that had a Centonics port, but not RS232, and disk
formats that the PC FDC couldn't handle. (such as TRS80 mdel 1, Sirius
9000, Vector Graphics, etc.), I wanted to connect their parallel printer
port to a PC, tell the alien machine to PRINT its word processor files,
and have the PC emulate a printer to the alien machine, but capture and
save what the alien machine "printed".
Unfortunately, I found that many of them had a very narrow, below spec
strobe. I wasn't able to make a tight enough loop on minimal PCs, so I
had to add hardware to extend the pulse. I had hoped that eventually,
people would have the bi-directional ports, and that all that I would have
to supply was software and a simple dumb cable, with appropriate pin
swapping FROM Centronics female to DB25 male, that as a dongle on the PC
would let any Centronics capable computer plug to the PC instead of a
printer.
And, at Comdex, I found it extremely difficult to explain the concept to
people who either had NO need for data from such machines, or refused to
understand that existing PC TO PC parallel port transfer programs used
special software, not just a printer driver, on the sending machine.
Since Laplink, etc. PUT their own software onto the other PC, why couldn't
they use Laplink to PUT the special software onto the Vector Graphic?
So, I concluded that I did not have a market for making the PC emulate a
printer for getting alien machines to PRINT to the PC, for disk formats
that the PC could not handle.
--
Grumpy Ol' Fred cisin at
xenosoft.com