A note on fixed price contracts.
Fixed price is *useless* without fixed *spec*.
This spec -> this cost.
Changed spec -> changed cost. Don't like it. Do without it.
It has also been demonstrated ad nauseum that retro-fitting something late into a design is *orders* of magnitude more expensive if it is not designed in from day 1 (or at least *provision* for it. EG the catapults on UK aircraft carriers so familiar to El Reg wartech readers)
Military projects are even *more* notorious than large software systems for being stupidly late and over budget. This is despite the fact that most of them are *physical* systems and not the *first* of their kinds, so there should be *extensive* data on relevant timescales with which to deliver estimates.
Techniques exist to cut cycle times (set minimal specs, freeze early, build early etc) but *none* of them seem to be used. Given how many military systems have to interface to existing systems I'm amazed that most of their design is not set *by* those interfaces, leaving just the core functionality to be specified.
And WTF does it *always* seem to be that *canceling* the project costs as *much* as running it to completion if not more? Where did that little "law of the universe" come from.
I note that UK Ministries *can* change. It used to be *gospel* at the Home Office that "Good times burglaries go up (more to steal) , bad times burglaries go up (more desperate people). New Labor seem to have managed to break that ideal.
It seems to have taken target policing, an increase in the number of people in prison and the *splitting* of the department into 2 pieces to do it.
I'm not sure what it would take to get that improvement with the MoD.