I hear what you are saying--if launches were easy they could have dropped a bunch of expensive constraints. But consider that if launch costs were truly the bot

I hear what you are saying--if launches were easy they could have dropped a bunch of expensive constraints. But consider that if launch costs were truly the bottleneck for JWST, they'd have just thrown more money at the launch budget and got a bigger / fancier rocket. I'm no rocket science person but I'd assume you can scale launch capacity by throwing more money at it--it's generally a solved problem whose risks are much better understood than whatever JWST is doing. It makes sense to throw money at things to reduce risk like larger launch vehicles, regardless of their cost. JWST is expensive because it does stuff that has never been done before. Lots of "unknown unknowns" that you need to make "known unknowns" Anything like this will be expensive--same is true for software even. You are working in uncharted water--risks are everywhere. If they could have dropped a bunch of their constraints and reduced program risk by throwing money at a fancier launcher, they'd have done it. The fact they didn't suggests that the cost & risk wasn't constrained by the launch. Now maybe you could argue that "if there was only a way to make it serviceable using cheap rockets"--then yeah maybe they could drop some of the more expensive and risky constraints. But however many years ago when this project started, such a thing was not even on the horizon. Only in the last few years has such a thing even begun to seem feasible. Perhaps future missions can relax their constraints because suddenly technology makes repair "easy". But JWST didn't emerge in that environment.