That is true, but, (if you haven't already), I would recommend looking into the delayed choice quantum eraser experiment. I think it is probably more true to the essences of what the article was trying to communicate.

I don't think I can do the idea justice, but the basic idea is such: if you 'tag' the photons in the double-slit experiment by which path they went through, then you necessarily break superposition and won't see an interference pattern.

However, if you 'throw' away this information completely, the interference pattern re-emerges. The delayed-choice part of it is just a way to through away the 'tag' information after the photon has moved through the slits. It seems to suggest that you can uncollapse a wavefunction if you just erase any knowledge you have of the measurement you made. (I've probably butchered it, read here for more[1])

Although, honestly, you don't need retrocausality, and personally I think it has much more interesting applications to information theory (how does one destroy information? How does the wavefunction know about our records? What even is information, man? etc)

[1] http://en.wikipedia.org/wiki/Delayed_choice_quantum_eraser