This text by Cliff Click raised my old thoughts on how future programming may look like.

I put it as comment to his text, and copy-past here as far.

Fancy, it seems working on Hotspot inevitably leads to this kind of thinking - mine started around 2006. My personal belief is that it's more fundamental problem than it seems, as we need to move away from the "recipe" style of programming, and look closer to the way reality and even our own computational devices organized. So far to me it looks like reality around may be treated not as consisting of objects, but could be naturally interpreted as field, with finite propagation time and quantum interactions. Such model naturally accounts for timing considerations, parallel operations, and exact stable machine states. But problem is that human education and thinking is centered around individual activities, "I go, I take book", which is mapped into the way we program (we essentially play little adventure game with our CPU and tell it what to do, and how to behave if it encounters particular monster or artifact). Moreover, engineers put this model of operations into the logical model of CPU.

All in all, my hopes are in area of "field languages", or more exactly "controllable field dynamics". Such languages need some theoretical work, but it looks worth looking at.