These two:
are effectively duplicates. The first asks a broad question including how to fake the effect in post-processing.
However, the answers to the first tend towards lens-based "real" approaches, whereas the second has answers focusing on software and post-processing.
The questions could be merged, but an alternate approach would be to make the first be focused on lens-based approaches and the second on software, and delete the answers in each one that no longer match.
Anyone have a strong opinion?