If it ain't broke, don't fix it (Score 1)

I'm back to working primarily with IBM mainframes. The shop I work in is still running code written in the 1960s, I kid you not... Much of that code is being slowly replaced (there aren't a lot of 370 assembler programmers floating around) with more modern code (database driven), but still on a mainframe. We have other types of hardware used for different functions, but the core of the business still runs on (smallish) big iron. If decades-old code is still meeting business requirements, it isn't easy to justify replacing it just because it isn't written in a "modern" language or runs on "cheap" hardware.