I asked this question of the fledgling Olin student product design firm IdeaTree at the end of the semester. "Why now? Why not 5 years from now?" In the case of IdeaTree, I was asking them whether it made sense to start a design firm now or to start working towards* a design firm with a firm deadline - no pun intended - of full operation 5 years in the future, when the students involved (all frosh and sophomores) have industry experience and contacts?

*This is what Chris and I are working on, starting off as "geekspeak translators" that help folks talk about their tech needs with engineers (usually through writing specs and giving technology overviews). We've had two projects so far and things are going well, although we really need a better way of describing ourselves and a punchier name than "Human Readable Specs."

Anyway. For any decision of whether to take an action or not, there's a list of Reasons Yes and Reasons No for any point in time - call these Y(t) and N(t) and label them green and red, respectively, in the figures that follow.

In the simplest case, you take the action at the maximum Y(t) and minimum N(t) possible.For instance, the question "when should I die?" usually yields the answer "as late as possible." (fig 1) However, "when should I wear diapers?" has a much different answer for most people (fig 2) and "biologically speaking, when should I have kids?" usually peaks in your twenties. (fig 3)

"But Mel, the max value of Y(t) and the min value of N(t) usually don't coincide." True. So you can take the difference Y(t)-N(t) (fig 4, purple) and go at the maximum value of that, since it's the time when the positives most outweigh the negatives.

"But every issue is both positive and negative!" (or: "for every good reason to do it there's also a bad one!") Okay, okay. If Y(t)-N(t) is always constant (fig 5), go for the minimum value of Y(t)+N(t). In other words, when you have the fewest issues and therefore the least situational complexity (fig 6, blue) which could lead to uncertainty and definitely just gives you More Stuff To Deal With.

Astute former ECS students of mine will notice these graphs have few labels, no scales, no units, titles, etc. This brings up one last point - that there's also the question of whether the question deserves as much thought as that - some things are most optimal when you don't bother obsessively optimizing them. There are more ways to deal with the world than enumerating and rank-ordering items - stop planning and start living. Do it now.