[tcarnell]

hmmm, has anybody studied reading patterns of people when reading from a screen? It may be possible to fairly accurately estimate the probability of a section of text being read.

For example if there are 6 comments on display at anyone time and they are on display (without scrolling away) for 2 minutes, one might arrive at a probability of how much of that page was read, and how much time was equally spent on previous/subsequent parts of the page.

Especially when factors like, people tend to scroll text to be at the top of the screen to read it, or like me, highlight text I am currently reading with the mouse. Using these factors, we might be able to know what is/isn't read...

before improving something, try to measure it first!

[bthomas]

Is there an easy way to capture scroll activity on a mac app? I'd love to analyze a data dump of scroll activity and active window.