There’s an aspect of automation which has always unsettled me: the binary nature of computers ultimately draws a stark yes or not line between two states. Take for example an admission exam. A single point is the difference between being admitted or not. There’s no space for the personal history of the student, and the closer one is to this line, the more it depends on luck. The student could have a rough night, a personal problem, or a simple distraction makes a world of difference in the outcome.

I used to think there was no automated way of distinguishing between these special case scenarios, perhaps anyone close to the cut-off line had to be reviewed by a human, but today I encountered this scenario. I wanted to be a little bit more clever about the dropping hints in the drag and drop mechanics, so if you moved a shape in the direction of a drop area, it would highlight that drop area.

However, I encountered scenarios where the cut-off area between one target and the next would cause a continuous flicker, the angle is right between this binary state, causing the target to jump from one place to the next:

Then I understood: it depends on momentum. If I drag towards a target, it must take into account not just my last data point, but the history of my data points to make an accurate prediction of where I’m intending to drop. The parallel with the admission exam is that an average student on a steady decline should be rejected over a below average student on an ascent. The single data point of an exam score tells us where the student was at the moment of examination, but it doesn’t tell us where he or she will be in the future.

I suppose I’m rediscovering statistical trends here, it’s just fun to find to figure out things and find parallels in one’s niche.