MutCell idea for collections

[clinuxrulz]

Just throwing it out there, and not expecting anything to be done.

I remember in an old language Smalltalk what we call methods today were once upon a time called ... messages . Can you believe that? Messages to me scream out the word Stream in my mind. And the message causes a mutation on the consealed state.

Well Stream::hold could be thought of as constructing a Cell where the message is the new state of the cell to take on after the current transaction. (After transaction to maintain consistency amongst all the snapshots)

With a bit of imagination we can construct the following.

Stream::holdMutator(sa: Stream<(a:A)=>void>, a: A): MutCell<A>

Which mutates the state after the current transaction. (Maintains consistency between snapshots in current transaction)

The benefit? Immediate performance boost for cells carry collection typed values.

[clinuxrulz]

There feels like a strong connection between OO programming and FRP.

In OOP we say what our wires go to.
In FRP we say where our wires come from.

FRP also gives us inversion of control for free. (The conciquence of saying where your wires come from.)

[cubuspl42]

I had exactly the same conclusion after reading the book. While, obviously, not exactly same, OOP and FRP are in my opinion much closer than authors suggest.