[tremols]

To ellaborate a bit further, I don't know haskell but I can agree that the authors code doesn't look like spreadsheet programming.

My approach is inspired by join calculus which is centered on the concept of waiting on a set of signals/cells to trigger a reaction (I ignore the concurrency capabilities of join calculus which make it a bit more complex). It seems to me that FRP takes the long route to achieve this since its focused on event handling.

Anyhow, this can be easily achieved with a simplified cactus stack where each stack "frame" holds a reference to its parent so that it notifies its parent when its value is set, then when all of the stack frames are ready the procedure associated to it is triggered.

This type of stack has some interesting capabilities including memory friendly infinite recursion and compound-key data structures. I believe that it has potential for artificial intelligence and the development of modern dataflow languages.