> > When I asked about making a parallel-port driver that was more low-level
> > a little while ago, I discovered that not all architectures can control
> > all the lines on their parallel port (the Amiga was an example of this).
> 
> That sounds odd to me. My understanding was that the Amiga parallel ports
> included some level of what was later added to the IBM PC as 'bi-directional'
> this sticks in my mind partly because there was parallel port network 
> software on the Amiga long before any such things on the PC. (DNet/ParNET)

The Amiga parallel port has some limitations.  The 8 data lines, select,
busy, and paper_out lines are fully bi-directional.  Strobe will generate
a strobe pulse whenever the parallel port data register is read or written.
The Ack line can't be read or written, it can only set a flag and cause an
interrupt to the CPU.  Err is not connected, auto is connected to Vcc and
Init comes from the system reset (can't be controlled under software).

There are add-on parallel port cards for the Amiga as well, and these may
or may not support the same features as the built-in port.

> It might be possible to make something work even for ports which can't
> manipulate some of the control lines, if there's a spare controllable line
> available, a custom cable solution could be made to work. Not the cleanest
> answer, but the hardware was only designed to talk to Intel boxes.

True.

  -jm