I agree the Wiki article is a little confusing when one is used to seeing t' = t / y. I tried to edit it to clarify a while back with the moving mirrors part, but my clarification ended up adding another mirror that didn't exist within the diagrams and probably complicated it more in the long run, so it was dropped. I might try again some other way at some point.
Anyway, the time dilation formula can be taken either way, only depending upon which frame we consider at rest, or which is our frame of reference. Usually we take the unprimed frame to be our frame of reference which is at rest while the unprimed frame is moving, but in the Wiki article, it is the other way around. It is still correct either way, however, since the primed frame is time dilating in the same way from the perspective of the unprimed as the unprimed frame is from the perspective of the primed frame.
In other words, there is no absolute "one frame is time dilating to the other", where either t = t' / y or t' = t / y, but each time dilates relative to the other in the same way, depending upon which frame's perspective we are taking. To see this, as mentioned before in another thread, we should really use four variables, say tAA, tAB, tBA, and tBB, not just the two t and t', since the dual purpose of the same two variables becomes unclear. With the four variables, tAA will be the time that frame A observes of frame A's own clock, tAB will be the time that frame A observes of frame B's clock, and so forth. Using these, we can now see that we actually have two sets of equations, depending upon which frame is doing the observing. For frame A, it is tAB = tAA / y. For frame B, it becomes tBA = tBB / y.