And of course nobody has satisfactorily solved the problem of drift (that I know of). That is to say, Trackpoints come with drift correction algorithms, but they don't always catch small drifts and sometimes they trigger when you're just making a long constant input, leading to a situation where releasing the input creates its opposite
That shouldn't be so hard to fix. Don't register any pressure if there's no finger touching the TrackPoint (determined by capacitance/pressure or whatever)--similar to how track pads have palm touch rejection.
