Here

Based on what I've read in Aviation Week, they have not decided if they'll use a modified set of Shuttle main engines or perhaps some other proven design like the Saturn J-2.

If the goal is to return to the moon and beyond, then using existing technology where appropriate is, IMO, the way to go. If the goal is to develop new technology, then change the architecture and dump tons of money on R&D. Being NASA, they'll somehow find the way to dump tons of money on R&D even if they reuse existing technology. Wise money management has never been a NASA strongsuit.

One thing I'd like to see changed would be a way to reuse more of the systems beyond just the CEV. They're throwing away way too much hardware. For example, I'd like to see a single stage lunar lander that could be refueled and resupplied before landing again. At first, I'd park it in lunar orbit following a mission and refuel/resupply it from Earth. If lunar resources do become available someday, then I'd refuel it from the lunar surface. I know you'd pay a pretty substancial weight penalty for this design, but given the lunar vacuum and 1/6th gravity, I believe it could be made to work. I'd like the lunar lander to be sized to carry a substancial payload to the lunar surface (say 10-20 metric tons) such as habitation modules. If the vehicle is sized to carry that payload down, then it'd have ample margin to return to lunar orbit without the payload. In an emergency, perhaps they could jettison the payload and abort back to orbit on one engine.