Yes :) it wasn't the clearest example. But is also slightly hard to express I think superscalar architecture for hardware benefits from an explicit graph of de

Yes :) it wasn't the clearest example. But is also slightly hard to express I think superscalar architecture for hardware benefits from an explicit graph of dependencies. Then equally the user benefits from reduced need to convert the stack to a graph with SWAP DUP etc -- which is one way to think of the role of those operators (as a stack to graph conversion) On the otherhand some problems are unrepresentable by a graph, and easily represented by a stack; for instance map, reduce, and a bunch of others What I'm suggesting is that both the hardware and the user and arguably even the problem, generally prefer graphs