An interesting variation of this was pointed out to me by a colleague while I worked in a very large company. We would often get requests from management to siz

An interesting variation of this was pointed out to me by a colleague while I worked in a very large company. We would often get requests from management to size various efforts. Often it would be hard to accurately guesstimate how long something would take. (Do we get to use a team of the companies most experienced kernel developers? Or will it be work done by some external company?) My colleague pointed out that often the management team had already made the decision and were just looking for numbers to justify their plan. They had a budget, but wouldn't reveal it. This meant that realistic sizings were often met with skepticism. "Do you really think that would take 22 person-years of effort?" Predictably, this led to the decision being made without the sizing (why did we have to spend two weeks on the estimate just to have it thrown out) or with an unrealistic adjustment made to the sizing (again why bother with the estimate). I hated the idea that the most experienced developers where treated like they weren't really onboard with the company's goals (because of the more conservative estimates), while crazy proposals for version two (with specifications for new features measuring 3 feet high) were thought to be genius; this led to version 2 being a year late and two-thirds of it's features jettisoned by the time of the release--along with the promotion of everyone that had made the worst mistakes. I worked at one company where the senior staff finally outlawed the posting of Dilbert cartoons on the development floors.