I find that the industry in general has a bias against gradual improvements of existing projects. Several times throughout my career, I've worked on mainline su

I find that the industry in general has a bias against gradual improvements of existing projects. Several times throughout my career, I've worked on mainline supposedly-legacy products while other teams work on greenfield rewrites, and the result is invariably that I'm able to improve the left-for-dead legacy project faster than the rewrite group can make something new --- and up with an objectively better product. The difficulty of continuity is overstated. It's very rare that something is so far gone --- even Fortran --- that it can't be improved by insensible degrees into something just as modern and usable as something started from scratch.