After reading about keyboard ghosting problems on the internet today, it seems that the kind of problem I was seeing is actually unlikely to be a ghosting problem.
As my co-worker Edward pointed out, ghosting due to matrix scanning shouldn’t be a timing dependent problem. If it is a ghosting problem, it should be perfectly reproducible regardless of timing, since it requires only that three appropriate keys be pressed down simultaneously.
Not only was the problem I was seeing not reliably reproducible.. it didn’t appear on some apps at all. This got me thinking… what else could it be? Then I rememebered that my keyboard and mouse are plugged in through a small bus-powered USB hub, which I had hooked up so that I could plug in the keyboard and mouse in one go. Could it be that my powerbook’s usb port just doesn’t have enough power to drive a logitech wireless mouse receiver, a the tactile pro, and a usb hub built into the tactile pro. This could surely be the case.
Unfortunately, based on further googling, apparently what usb devices report as their power usage, and actual power usage is not necessarily correlated. Devices can report whatever they want, and its up to their design to stay withing their limits. So I guess I will not really have any way of knowing except to try it without the hub, or with a powered hub for a while.
So long story short, it’s back out of its box and plugged directly into one of the usb ports to see if that helps. So far I haven’t been able to reproduce the problem that I was seeing the xterm so frequently. So far so good.