Unique Puzzle Solutions

When playing sudoku or similar sorts of logic puzzles, I occasionally hit a situation which can be solved if you assume the solution is unique. And it always bothers me as to whether it is a legitimate technique or not (only I could lose sleep over something so dumb, eh?).

Consider the partially solved sudoku puzzle to the left. Let me draw your attention to the squares I have carefully labeled 'a' and 'b'. By elimination in the center 3x3, these squares must either be a=7 b=8; or vice-versa, a=8 b=7. Meanwhile, looking at its row and column, 'c' must also be either 7 or 8. But looking at rows, columns, and 3x3's, 'd' can be 6, 7, 8, or 9.

However, if we know the solution is unique, we can eliminate 7 and 8 as possibilities for 'd'. For example, imagine hypothetically that the "unique" solution had d=7. Then c=8, b=8, a=7. Well, then if we replace that "unique" solution with d=8, c=7, b=7, a=8, you would have another valid solution. And therefore our supposed "unique" solution wasn't so unique, so you can assume 'd' is either 6 or 9, and the empty square between 'b' and 'd' therefore must be 7 (question for the reader: why?) and the rest of the puzzle just sort of falls into place.