This is a really nice idea. Because actually, I think the skill you actually want to test for with this is not so much working with incomplete specifications, as much as recognizing them. I've known programmers who would get a project, think they understood what was required, and only after coding for a month realize there was some unspecified but very crucial detail, and how it actually needed to be worked out ended up invalidating a large chunk of what they'd already done.
in reply to Re: Re: Re^2: How to measure Perl skills?
in thread How to measure Perl skills?
I guess the best way to get real world ambiguity would be to ask marketing to write up the test problems :-)