@Hans
RE CNC - I can see your point as a hobbyist idea (it _could_ be done) but what kind of CNC are you considering that needs such low latency feedback from the sensors to control stepping motor functions?
My comment about feedback control loops wasn't really about CNC. That was in response to the "you can't do real-time" comments from Thorham. Like you said, you usually don't really need such control with stepper motors. There are a few cases where it might be needed. For example, you might be machining something that need your tool to apply constant pressure.
Now, I don't have a specific application in mind beyond a desktop milling machine. I bought the motors before the A1-X1000 came out, and my plan back then was to get the computer to turn the g-code into a set of stepper motor bit-streams, and upload that via wifi to a small ARM controller board. That board would store it on an SD-card, and spit it out to the motors once done. However, using Xorro would be faster and more convenient.
I genuinely don't see what advantage Xena/Xorro bring here, and do see disadvantages of complication. This doesn't seem to utilise Xena/Xorro much beyond (as you have observed) a parallell port.
I could drive more outputs via Xorro and read more inputs than a parallel port could. This is one of those situations where you need all motors to be perfectly synchronised, so using multiple parallel ports would
not work. This is irrelevant for a 3-axis machine, but...
Another consideration is speed. With stepper motors there's a trade-off between precision and speed. If you want both high precision and reasonable speed, then you'll need to be able to generate a very high-speed pulse stream (and have motors that can keep up). I don't know what the switching limits are when you bit-bang the parallel port, nor have I checked what the speed limits of my motors are. However, I'm pretty sure that I could get Xena/Xorro to deliver a faster and more precisely timed set of pulse-streams than a parallel port could, without bringing multitasking to a halt.
Finally, let me ask you this:
Perhpas something like robotics control maybe - considering a 'walking robot' that requires constant feedback for eg balance.
However like any specialist application, it's better to choose the most appropriate solution to the problem, not start with a given solution and try to fit the problem around it.
I remain confused as to what function Xena/Xorro provides that is advantageous.
I've made a few suggestions above, although I fully expect a "yes but..." reply. However, does it really matter if my CNC controller uses Xena/Xorro and the PA6T to their maximum potential? I don't think so. So much hardware and software doesn't (e.g., just think about how many games you play actually uses your computer to its full potential).
Similarly, there are myriads of microcontroller boards available out there, and you're always going to find a different board that is "at least as good" as the one that you chose. Is that any reason to cancel your project? No, it isn't.
You are too fixated on "where's the killer advantage" to even begin thinking about what you could do with it.
In the end, we have a CPU with high processing power coupled (via a low-latency interface) to a device that excels in things such as custom application-specific I/O, motor control, etc. I find that idea interesting. Plus, I like seeing what I can make things do, irrespective of what other people say it's predefined purpose is (e.g., 3D graphics with 2D compositing).
I really hope that people will think a bit more creatively about what they could use it for, and be less quick to look for reasons why those ideas are "stupid." You might just find a few good/interesting ideas amongst the "silly" ones.
Hans