Quite possible the trackpoint was genuinely worse in 2007 than in 2001. Like Windows 95 itself, huge amounts of human factors research went into making it worthwhile when it was new and unproven, with usability studies deriving the optimal curves and constants. People described it as feeling like telepathy. And like Windows itself, some time between then and now, everyone seems to have forgotten that it's a
hard problem and just started YOLOing whatever random stuff "feels right".
Rant: On my T480 - an extremely popular, flagship Thinkpad - the trackpoint is unusable out of the box on Linux - it crawls at a snails pace unless I manually max out the sensitivity in sysfs, and also fiddle with xinput (no configuration is possible on Wayland). What is worse, infuriatingly, the curve seems to frequently change whenever libinput is updated, requiring me to go and find another set of fiddles to bring it roughly into line. This is clearly not the output of a structured, empirical, research-based approach to the issue, but of someone just fucking around with the constants for one reason or another. It does not feel like telepathy, and it makes my wrist hurt in short order if I overuse it (luckily the laptop also has a generous touchpad).
I dunno, maybe it's better on Windows. But I'd be amazed if anyone cared enough to make sure that 90s usability research was still valid, on modern variants of trackpoint sensors and modern high resolution screens.
