News that Granger is abandoning the sinking ship that is NHS Connecting for Health is hardly a surprise. The £12 billion project was doomed before it started, because it tried to apply an unworkable, 20th-century, closed-source software methodology - one that not only does not scale, but that actually gets worse the bigger the project (hello, Fred Brooks).
The only way to address these kind of mammoth undertakings is by using a lightly-coupled, decentralised approach. And that means open standards at a minimum, and ideally full-bore open source. The equation is simple: the more openness, the greater the scope for componentisation, the greater the flexibility - and the greater the chance the damn thing will actually work.
Sadly, NHS Connecting for Health will go down in history as the perfect demonstration of this fact. - Sadly, because I shall be paying for some of it.