Lt Col. Dan Ward (USAF) from the Defense Acquisition University presented an interesting acquisition lesson from an unusual source: Star Wars. His article, Don't Come to the Dark Side (PDF), highlights why "Death Star"-like projects are a bad idea, for both operational and programmatic reasons.
The article acknowledges, but skips over, the standard fare of "Darth Vader as leader." ("I don't think we really want PMs to walk around in capes and black armor ... his path is not one we should follow. I'm pretty sure it leads to suffering.") Instead, Ward talks mostly about why a project the size of the Death Star was a big mistake:
The truth is, Death Stars are about as practical as a metal bikini. Sure, they look cool, but they aren't very sensible. Specifically, Death Stars can't possibly be built on time or on budget, require pathological leadership styles and, as we've noted, keep getting blown up. Also, nobody can build enough of them to make a real difference in the field.
The bottom line: Death Stars are unaffordable. Whether we're talking about a fictional galaxy far, far away or the all too real conditions here on Planet Earth, a Death Star program will cost more than it is worth. The investment on this scale is unsustainable and is completely lost when a wamp-rat-hunting farmboy takes a lucky shot. When one station represents the entire fleet (or even 5 percent of the fleet), we've put too many eggs in that basket and are well on our way to failing someone for the last time.
The answer isn't to build more, partly because we can't and partly because the underlying concept is so critically flawed. Instead of building Death Stars, we should imitate the most successful technology in the saga: R2-D2.
Ward concludes that, rather than sinking your energy into huge, expensive projects like a Death Star, you should instead focus on smaller, more reliable systems like astromech droids:
The key is exercising design restraint, focusing our requirements on the essential requirements rather than the endless list of desirements, living within our budget and resisting the temptation to extend the schedule. [...]
There are all sorts of ways to simplify a design, to reduce a set of requirements to the bare minimum, to make sure we build what we can afford. Don't believe such a thing can be done? That is why you fail. But those who do believe will find the system they built just might be "the hero of the whole thing."