We’re stuck with a serial life cycle. We have very strict phase gates and have a management review for each of them. But in between the phase gates, we use timeboxes and iterate on prototypes to know how good our estimates are.
We’re in the design phase. And we realize that we can’t get the performance we need out of that particular component. It’s not going to happen; it violates the laws of physics. It’s time to replan what we can do and release a slightly different product.
I’m not worried about the schedule. I plan in detail for only four weeks at a time. I’ll have to throw out part of my planning for the ...