Gregory's Server
@raphael I think the answer really depends on how bleak your outlook for the future is. To some degree permacomputing is speculative computer science, so (in my opinion) really it's a question of what problems you see actually coming to pass that would require making changes to how you write software today to ensure that software is still useful in the future.
@raphael for example, I think my own outlook is fairly optimistic in terms of availability of computing hardware and energy availability, but I'm pessimistic about the digital divide and expect we will have more prolonged major supply shortages here and there in the future, so in my own projects I put a great deal of thought into keeping minspecs as low as possible, portability, and avoiding dependence on bleeding edge hardware features