But you have to think of the ethics of it. In this case, the materials to work from are an incomplete workprint, a pan & scan DVD of the first release version, a widescreen DVD of the further cut U.S. release, the recut workprint, and a few odds and ends.
On one level, I think that it should be as close to the workprint as possible - nothing less and nothing more, warts and all...
There's a bigger problem, though... the point is to get people interested in the film so there's an incentive for Disney to put the money into a true restoration. One problem I've seen with some "scholarly" restorations is that they didn't really take any precaution to make sure the film is easier to watch. That's why we're getting Calvert animation in place of storyboards - as mediocre the animation is, it still grabs attention better than the storyboards (which are superb, anyways). There's also the factor of having to use audio from the recuts because of higher fidelity (or in same cases, due to the workprints having unusable audio).
We can't assume the workprint is the exact flow that Richard Williams intended - mainly due to missing footage that doesn't turn up until the Calvert workprint, unfinished audio, and the screenplay being so loosely followed. For example, we know that the shot of the Thief trying to steal an emerald was much longer since footage appears ONLY in the first recut's credits - same thing for the attempt to use springs to get the golden statue. We can maybe assume Williams meant to cut it, but the workprint still has a shot of the Thief walking away injured with a spring on his foot. Then, we know that Williams intended to trim parts of the Thief pole vaulting and the war machine stuff. But what would he have cut out?