I think Ted has it closer than Bill, although I can see why Bill says that. I once heard someone describe it as "making a player hit all the little shots that he really should" in order to make a better score. Ideally, the variety of shots aspect is important in achieving the GD mantra of rewarding length, finesse and accuracy as equally as possible, and allowing all types of players to generally be competitive.
I have also heard it desrcibed (by pros) as that, AND as having the course set up for those shots, such as asking a player to hit a high spin shot on a typically downwind hole being bad shot values, while asking for high spin into the wind would be good shot values, because nature/physics helps make a good shot achievable, rather than next to impossible.
Other attempts at shot values include the USGA rating system, which starts with length giving a pre assigned value to a hole based on length. The rating can be affected a maximum of 0.1 stroke per hole based on a combination of factors like target size, hazards, etc. The slope rating is similar, using about a dozen factors to alter the rating further for the 20 handicapper.
Hurdzan discussed this in his first book and made a good point. If you size all greens according to the USGA chart (about 15% of approach shot length wide and 22% approach shot lenght long - i.e. for a 160 yard shot, 24 yards wide and 36 yards deep) then don't all approach shots have the same degree of difficulty? Some should be smaller than "prescribed" and others should be larger for a variety of challenges.
Of course, variation in holes of similar length and par should be figured in - long par 4's don't need to have wider fw and bigger greens, even if the USGA formulas for rating suggest it. If there were four long par 4's, I would think:
One would have a wider fw and smaller green,
One with a more difficult tee shot and easier green target. One with more difficulty on both shots and easier putting.
One where par is defended by tricky putting.
Would have better shot values than one where they were all similar because of some formula.
Thompson did his shot values chart, which emphasized approach shot length variety, (i.e. use every club in the bag) and others have done "wind compasses" to assist in determining the variety of holes relative to wind direction (and usually par - with all par 3's and 5's ideally pointing to different directions, for example.) To me, length alone isn't enough and I trust prevailing winds about as much as I would a riverboat gambler.
If you took all of this together, I guess shot values would encompass variety of all kinds (wind direction, length, shot pattern, etc), overall shot challenge (shot type favored) vs. difficulty of shot (target size and penalty), while putting them "all in the right type of holes and right places in the round."
Let the debate begin!