Computer scientists and time

Computer scientists have a very interesting relationship to time. Obviously time is a critical component of many applications of computing, whether it’s the running time of algorithms or computing behavioral models based on observed actions over time. And yet computer scientists themselves don’t seem to be very good at time.

Case in point: the computer science “5 minutes”. Computer scientists seem to have an unshakable faith the almost any discrete computing task can be done in 5 minutes. When I was a grad student it was a running joke that you could tell which significant others of lab mates were new because they actually believed the person they were dating when they called up and asked “Are you almost ready to go?” and got the response “Sure, I’ll be ready in 5 minutes”. The significant other would then show up at the lab and spend the next 30 minutes twiddling their thumbs waiting for person to actually be done. More experienced significant others would call and then just wait 30 minutes before actually showing up.

The computer science “5 minutes” occurs for almost any task that can be associated with a “Are you ready?” or “Are you done?” question. “Are you ready?” “Sure, let me just finish adding this feature. It’ll just take 5 minutes.” or “Sure, let me fix this one last bug. No more than 5 minutes.” I think it’d be an amusing CHI paper to study whether computer scientists just have overwhelming faith in their programming abilities (It’s not hard! Surely I can finish it in 5 minutes.) or whether time just disappears while their coding, such that they lack an appreciation of how long programming tasks really take. I suspect the latter.

Another common examples ties to our ability to manage time. If you ask a computer scientist to perform a task with a short term deadline (say, this week or next week), the answer will almost always be no: we inevitably have too many things on our plate and just can’t handle one more. So how do you get a computer scientist to take on a large, time-consuming task? Easy: ask them 3 months in advance. We appear to have a limitless faith that somehow in the future we’ll have much more free time. Why we retain this belief in the face of overwhelming evidence is unclear. Perhaps we somehow expect that the projects we wind down will not be replaced by new projects, despite the fact that has never in our lives happened. Or perhaps we expect that in the future days will somehow contain more hours or hours will get slightly longer. Or maybe our faster computers and better user interfaces will make us that much more efficient.

Regardless of the cause, if you ever need a favor from a computer scientist that involves an investment in time, just ask for it 3+ months in advance. And never believe a computer scientist when they say they’ll finish what they’re working on in 5 minutes.