How many companies realize the true cost of failing to plan for technology migrations? The death march is not just a kind of failing project, it is also the character of an entire organization when managed with non-lean methods.
As time goes by, the technology terrain shifts and is sculpted by forces of (a) academic transfers, especially but not exclusively of STEM research, (b) open source and proprietary developments, including those of standards organizations and (c) government and quasi-government fiat. Organizations such as Orange UK attempting to set a fixed goal fail before they've deployed a single line of code, because the target itself is moving, but the space it is in is also growing, shrinking, skewing, or otherwise being subtly or suddenly transformed. Not to adjust the sights once in a while is a blatant management failure, not a mistake.