I think if you'll search the archive you'll find we hashed over the idea of building all new locomotive pretty thoroughly early in the winter (December? I Noremember...). In a nutshell the big expense is a new boiler, there is no way to cut corners without cutting safety. As the excellent posts from this great bunch of knowlegable people pointed out, much of the rest of it would depend on how many concessions you were willing to make to modern technology. Making use of modern production methods (CAM, EDM, lost foam casting, laser and water cutting, complex weldments, etc.), and modern metallurgy, you could bring the build costs way down, but at the trade off of not having an exact replica. There could be additional savings by building several identical units at a time, but finding buyers might be tricky.
I believe the merits of narrow gauging a new(er) chinese lokey was suggested, and discussed around that time as well.