That is actually kind of interesting, because I have a diametrically opposing view, equally fierce, of this (unless we are misunderstanding one another).
While I agree that the interpretation of an ESTOP *software* signal should be in tight coordination with LinuxCNC (I don't really see that a pendant should not not be able to inform LinuxCNC of the desire to enter that state), I am quite convinced that any true EStop should have an electrical interrupt of whatever allows a machine to move (axis and spindle). The best would be power to the responsible amplifiers, but forcing a disablement within those devices may suffice. This is a failsafe that is completely independent of the computer, which is an important and intentional condition.
To me, the estop button on the pendant is the same as any other estop button I may add to the machine. All are wired in series so they all act the same (and, intentionally, a broken wire or disconnected pendant, will result in an estop state). The fact that Vista has two poles on their estop, one of which is wired into the chain (above) and the other is interpreted is a little bit of a curiosity to me. I'm less of a purist about the software interpretation of the switch. I was, and frankly still am, kind of on the fence about that one.
Anyhow, that is just my view. I am always interested in what others feel, and why, because it helps me decide if what I'm doing is still correct (in my mind). Your views, and those of Andy's earlier in the thread, are all welcomed and appreciated. They are thought provoking and that is a good thing. Let's face it... LinuxCNC isn't for the feint of heart. It is for people that want to invest in learning and, hopefully, are open to different ideas that challenge our own. All good stuff.
