4 ms·
IO solves the "time problem" by just encoding a von Neumann machine in the functional language, which is less like a liberation than a peace treaty.
by edflsafoiewq 5y ago
IO solves the "time problem" by just encoding a von Neumann machine in the functional language, which is less like a liberation than a peace treaty.
- creata 5y agoMaybe my imagination is too small, but what else could a solution look like? You want sequencing, and IO gives you exactly that.
- dustingetz 5y agostream, signal, and differential programming
- mumblemumble 5y agoPerhaps. But I think that you're also wandering away from the plot. Clinging to our contemporary way of thinking about this stuff isn't really going to help us understand what he was grappling with 30, 40 years ago. It's more likely to muddy the waters. I don't see any evidence in there that Backus was looking for anything quite so lofty as the stuff you're bringing up. While it's true he wrote the paper that kicked off the whole idea behind FP, we've also got to keep in mind that, as he repeatedly mentions in the interview, he was a lifelong Fortran programmer who, in marked contrast to most fans of FP nowadays, held a lifelong affinity for Fortran. It sounds to me like encoding Von Neumann bits into the functional language so that you're there when you need them is exactly the sort of thing he was looking for. To my reading, his Turing lecture would seem to suggest the same.