I think we can agree to disagree on whether or not to run a new cable as opposed to using the pport for the additional I/O.
Putting that aside, Windows is not a real time operating system, and neither is the ethernet port, which is why external motion controlers are better than the pport for motion control.
The data from the ethernet port has to be buffered and sent to the dspmc, and is not handled in real time.
When using a mpg, you are generating pulses at a high speed and you could and probably will exceed the speed at which the pulses can be accepted and handled by the pport and then sent to the dspmc.
I would say as long as you tell the operator that the mpg can't be trusted and he uses the dro's as gospel, you would be ok. What I mean by that is that if you are setting up and want to move say 1" and you crank the wheel ten times assuming .100" per rev and the machine only moves .920 or so, it will not engeder trust in the controls that you have sold the customer. While most of us would watch the screen and not count the MPG revs, when the 2 don't agree at all times it will make people think, if this doesn't work properly, what else doesn't work properly.
I think that things like coolant on/off, spindle control, and other non time sensitive stuff would be fine. E-stop, feed hold, limit switches and anything else time or safety critical should just be wired thru the 7711 to the dspmc.
Just my $.02.
Mike