The US procurement problem for big ticket ships and airplanes is a vicious circle. We don't get to buy many platforms because they are too expensive. So we try to pack every single capability we ever wanted onto each platform, which makes the platform too expensive to buy in quantity. This leaves us with fewer platforms. And repeat as long as possible.
I'm a USAF Vietnam Era veteran, so I mainly watch airplane development. I would have to call the F-35 the second coming of the F-111. The F-111 was a Vietnam era fiasco. After a lot of expensive bureaucratic bumbling, it ended up in service only with the Air Force
Just like the F-111, the F-35 is a multi-service airplane that costs too much and requires 3 services to sign off on design changes. That many admirals and generals with their fingers in the pie just has to be prone to management delays, which delay the project. The F-35 is a jack of all trades, master of none. It should have been developed for one service, probably the Air Force, before any of the other services had any input. After IOC, the Navy and Marines should have had their turn to modify the design into something they could use.
But I think the biggest problem with US development of big ticket items like ships and airplanes is "systems thinking." We design and develop new everything for most new designs. Not just a new platform and engines, but new electronics, new software, new ejection seats, new cockpit displays and, in the case of the F-35, a new on board logistics system, which isn't working right now. Having all of these new things increases risk tremendously. In integration testing, there are an exponential number of combinations of new elements which may not be working together correctly.
I think the solution is to separate the platform development from the electronics and the software. Develop the new platform with as much off the shelf electronics and existing other standard parts as possible. Don't pick a completely new part unless you have to. This makes debugging problems in the new platform a lot easier. It's either the software specific for the new platform or it's the platform itself.
Electronics and revolutionary new software should be deployed after IOC, Initial Operating Capability. We definitely know how to do this. We upgrade existing platforms routinely. Those B-52's on missions over Syria are not using vacuum tube electronics from when they were originally built in the early 1960's. Even the A-10 Warhogs have upgraded electronics and the Air Force doesn't even like them. New weapons systems, like the small diameter bomb, can also be integrated after IOC.
I'm sure the Navy can do this with ships. When the battleships sunk at Pearl Harbor were repaired and refitted for service in World War II, their secondary armaments were changed to feature a lot more anti-aircraft guns and they also got fitted with radar. There’s no reason similar upgrades can’t be fitted into modern ships after their Initial Operating Capability.
It’s a lot cheaper to go from a platform that works to an improved platform that works than it is to debug a completely new system over the course of several years. If it takes several extra years to debug a new ship or airplane, you get no benefit from your investment while you’re debugging. The platform is useless.
The whole idea is to get something that works ASAP, then improve it with additional capabilities or variants. This is the way World War II aircraft were developed. The P-40 was an adaptation of the P-36 with a better engine. The P-51A used the same engine as the P-40. The subsequent P-51s used the same engine as the British Spitfire.
Similar component sharing and variants are possible today. For example, there was a proposal to build a stretched F-22 as a bomber.
Starting with off the shelf capability with everything but the platform and engines will make it harder to allow scope creep, the repeated inclusion of “just one more requirement.” If only the platform and engines are developmental items, it's not feasible to make a Swiss Army plane or ship that’s designed to do everything. This will hold down costs. You have to tell the people asking for additional features that it will be fixed in the next release.