Directionality
Discussion thread | Status |
---|---|
Drafting as of Oct 25, 2016 |
The following contexts exist in a hypothetical application:
A B C
Our application starts at context A. The user then performs the following transitions:
- A to B.
- B to C.
- C to A.
There are two ways to think about how we move between contexts: from/to and back/fore.
Traditional from/to transitions
In the traditional from/to model of transitions, the above context changes would have the following variable values:
Transition # | from |
to |
---|---|---|
1. | A |
B |
2. | B |
C |
3. | C |
B |
We're concerned about three distinct transitions and will write three distinct code paths.
If we wanted B => C
to fade C
in and C => B
to fade C
out, we might try to write one function that looks like so:
let animation = Tween("opacity", duration: transition.duration)
if transition.initialDirection == .forward {
animation.from = 0
animation.to = 1
addPlan(animation, to: toViewController.view)
} else {
animation.from = 1
animation.to = 0
addPlan(animation, to: fromViewController.view)
}
Note that we have to check initialDirection
in order to determine which view to add the fade plan to.
In practice, the B/C and C/B transitions are often mirror images of one another. Is there a better way to think about transitions with this in mind?
back/fore transitions
In a back/fore transition, the above context changes would look like so:
Transition # | initialDirection |
back |
fore |
---|---|---|---|
1. | forward | A |
B |
2. | forward | B |
C |
3. | backward | B |
C |
Note that our back
and fore
variables now has just two distinct permutations. If we were to write the B/C transitions with these variables our code might look like so:
let animation = TweenBetween("opacity",
window: transition.window,
segment: .entire,
back: 0
fore: 1)
addPlan(animation, to: transition.foreViewController.view)
transition.foreViewController
is a fixed value, regardless of whether we're transitioning forward or backward. This eliminates the need to express an entire logical path of code.
Context transitions use the back
/fore
model in order to minimize the amount of code required to express bi-directional transitions.