Understand this. The F-35 design principles are very heavily reliant upon a lot (no, a LOT) of complex, immature, expensive and (in some respects) delicate technologies.
The same challenges and risks presented themselves to the Comanche program, with some bleeding edge tech, and some significant software challenges, none of which were mature. A variety of things led to that program being cancelled, but cost was a significant factor. Even in the mid 1990's, upper level officials (in an era of austere budget, Clinton era) seriously considered killing Comanche and funding other programs. It took another half a decade for the axe to actually fall.
The difference I see in JSF is that there isn't overlapping capability projected for the outyears. It appears to be "she's my only choice for a date, so I will be dancing with the fat girl. I want to dance at the prom."