I think this is a very nice start at pulling the landscape together.  I note several comments from email exchanges (due to initial Gridforge config 
problem not allowing comments - now fixed so comment away):

- because the draft tries to look down the road in addition to near-term "low-hanging fruit" there was discussion about the focus, and I believe 
confirmation that we do indeed want to try to gain interoperation between operational grids, and it's not cheating if we leverage the fact that we are
 running the same software.

- informed by this effort we will be in better shape down the road when we look at using things like JSDL to interoperate even without the same 
software.

But it is important in this business that we avoid the pitfall of "wait just a few more months, really this new thing is *almost* ready and it will be
 so much better" because there is no escape from that loop.

Speaking for TeraGrid, I am eager to try to interoperate in the very near term, and no fruit hangs too low for my tastes.  

(and I think this draft is a good start!)

CeC