Thank you, thank you! I did wonder: "how hard can it be?" and "what could these libraries possibly do?", but I said to myself: "Martin, don't re-invent the wheel". You confirm my initial hunch. So, but where to go from where I am now? I hear the message, but I don't know what to do. Am 03/17/2014 11:02 AM, schrieb Sacha Sokoloski:
So I'm also doing a lot of simulations, and have experimented with FRP as a basic toolset for defining my simulations.
I experimented with Netwire for a while, and while I still think it's a good library, I eventually threw it away to rely simply on Mealy Arrows (Netwire is Arrow based FRP).
From what I've seen, the point of FRP libraries is to handle interactivity. That's what all the functions in the libraries are about. If you do have a virtual time stream that you can simply define at runtime and doesn't require side effects, you may find, as I did, that you'll create a simplified type synonym for the FRP structure that you're working with, and then not really using the libraries at all. The point is ultimately that, although pure simulations could nicely be part of any more complicated FRP program, if all you want is to do pure simulations, then you'll be introducing a fair bit of computational/structural overhead to fit it within the FRP framework, without any real payoff.