I thought this was a decent article, with decent advice if you're in a situation where you have to make estimates under conditions of uncertainty. An example I

I thought this was a decent article, with decent advice if you're in a situation where you have to make estimates under conditions of uncertainty. An example I've tried to use to explain software deadlines is the college term paper vs the mathematical proof. I can commit to writing a college term paper by a set date. I may do a good job or a bad job, but I know I can produce 20 pages on a topic with citations and references. It's really just a matter of will and follow through. Many people experience deadlines this way, which is why they get a bit outraged when software developers fail to hit deadlines or warn properly that they won't. They don't understand that software can be more like a mathematical proof. You can tackle it, try things, but you only might crack it. You might be no closer than when you started. You might be moments away from cracking it and not knowing it. My career advice to people is to seek out situations where software development deadlines have more in common with term papers than mathematical proof. These jobs do exist, though they are elusive. It's just another desirable aspect of a job, like better pay, nicer working conditions, telecommuting. There are jobs that define the goals of a software project more vaguely, to the point where you can deliver something great, or something merely ok, but there's really no chance you can't deliver something at all. Some jobs go extinct because they are just so unpleasant that the people with the talent to work them simply find other options for employment. I personally find the stress or working under strict deadlines with very little certainty unpleasant enough that I'll accept lower pay or other tradeoffs to avoid them.