Indeed the mechanism is very deliberate.
Moreover, the present sample is the second version of the puzzle.
Is version 2 the same puzzle with the same solution just with a different outward appearance? I'm assuming it may be possible to build similiar puzzles with different solutions... not sure if that's true or not.
Yes and no.
I compared the 3D view at Shapeways with my pieces and it's clearly the same combination of parts but with a completely new dovetail mechanism.
But even with the same mechanical behavior, the puzzle will solve differently because the use of color is different. I own two of the old-style Turn Apart puzzles and they're surprisingly different to solve. I've had the colored one apart several times because I can't get it it together without two identical colors touching. (I know it's possible.) In contrast, the all-white one poses the difficulty that the pieces are externally indistinguishable so it's hard to know what moves I can make.
Turns Aparts.jpg [ 277.54 KiB | Viewed 1011 times ]
So it's easy to see how the newest model makes it even easier to indentify specific parts (and hence know what moves they can make) and yet much harder to solve because each part has only one solved position. The all-white Turn Apart had 240 solved positions, the old-style colored model had 40 (or fewer) solved positions if matching colors are disallowed, whereas the new model has only 1 solution in any case.
Just as the 2x2 and Pyramorphinx are different puzzles despite using the same mechanism, each Turn Apart model offers a different solving experience.