I know of at least one semiconductor company pushing the state of the art on 1990s litho processes, running 1990s control software on their stepper that runs on

I know of at least one semiconductor company pushing the state of the art on 1990s litho processes, running 1990s control software on their stepper that runs on Solaris on SPARC32 (SPARcstation 5). Anecdotal, but their production engineers have stated that not relearning how to use critical parts of their process every few years has helped them stay focused on actual R&D. This is of course an extreme, probably unusual example.