[Silhouette]

It's a poor choice of scenario for an interview, because no-one who knew anything about writing chess software would ever write that sort of OO code. If the point was to test for an understanding of OO design, wouldn't it be a good idea to provide a problem where OO design was appropriate?

[fjsolwmv]

OO is for large problems, not 30-line whiteboard able problems. That's the challenge.

[Silhouette]

Is there any reason to believe that code using efficient data representations like bitboards can't scale just as well, if not better, than OO versions full of classes?

Or that the interface to use that code would be harder to understand or maintain?

If not, why mandate the use of OO at all, and if it's OO you want to test, why not pick a better example where it might actually make sense?