Bob Armstrong wrote:
Jerome Fine
(jhfinedp3k at compsys.to) wrote:
How would device drivers designed for a Qbus or Unibus do disk I/O?
How would serial ports to a terminal operate?
This is the killer with the S-100 plan. I've got a bag of T-11 chips and
had considered doing something like the SBC6120 for a PDP-11, but it's
pointless unless all the peripherals are also PDP-11 compatible. Unless, of
course, you have lots of time on your hands and want to write your own
PDP-11 OS and tools, and that'd take an order of magnitude more time than
designing the hardware.
You could, if you had the right bits and knew enough about RT11, just
write new device drivers for RT11 and port it to S100 hardware. I'm sure
there are people with enough knowledge to do that, but RT11 is _not free_.
What hobbyists do in their garage is one thing, but any commercial endeavor
would have to be careful about putting their name on that.
AFAIK there are no "free" OSes that will run on a T11. 2bsd is your only
option, and that'll need at least an F11 CPU. An F11 S-100 card would be a
cool project (I'd buy one!) but it'd be much harder. F11s are a multi-chip
chipset, much less well documented at the hardware pin level, and less
available than T11s.
There's always the FPGA route, but then the S100 becomes superfluous. You
could fit a whole 11/73, including all the peripherals, into any modern FPGA
without even trying and all you'd need beyond that are the connectors and
level shifters.
But then, what do you do with all that hardware once
you get it working? I realize that the fun for hardware
addicts (as opposed to software addicts like myself)
is to get the system working and running the OS.
Is that the end of the process? With software, there
always seems to be a bug to be fixed or, especially
interesting and even more challenging, an enhancement
to design, implement and debug. I remember when
I wrote some code for a program that made the program
Y3K ready. The OS still does not have the required request
to process date information for after 2099 CE, but now
the code is there waiting. However, when I was testing
the program under an earlier version of the OS, the
program crashed. I finally figured out that the Y3K
ready requests were handled (actually - and correctly -
rejected by the OS) by the current (somewhat
recent - only 20 years old) version of the OS, but OS
versions that were 25 years old could not handle the
Y3K requests - which were not yet in any OS.
So those challenges always exist with software. What
about hardware after the hardware is finished?
Jerome Fine