For ductile materials the von Mises theory gives very good results as a criterion to compare the stress due to simultaneous traction/compression on several directions and shear strains with the linear traction test done on test machines. It is a measure for the local level of strain in the part. It is broader as validity than the Tresca criterion.
I have thought about this for awhile now and have decided to pose these questions back to you (so I don't waste breathe/electrons)
What exactly are you looking for? How much detail? Do you have access to the two different codes 2004 & 2007?
Are you just looking specifically for the changes to the design stress basis? Because there were also a lot of other changes (it was completely rewritten).
In a broad sense the result of changing the theory of failure from the maximum shear stress to the maximum distortion energy (von mises theory) - is a more accurate analysis which leads to calculated stresses closer to the real developed stresses. This allows for a less conservative design and a savings in material and weight.
Take, for example, this situation which was resulting from the beta testing of the new standard: