- I like to use my tele lens with a lens2scope for direct viewing. It's fantastic what image quality even the 100-400ii gives. the new superteles can not be used this way, because they can not be focused without camera. Canon should offer a lens2scope themselves, which can activate the IS system (with included battery of course) I would preorder it just now.
- Handholdable is relative, people are very different in strength. One finds a 70-200 2.8 not handholdable for a longer time, the other can manage a much heavier combination.
- The price is almost 2x the version ii. it's one very expensively saved kilo (even if the balance may be better on top
- In the TDP measurements one can find more inconsistent results, if one compares them to the MTF lines, especially in the super teles. So, the worse center result with 1 of the extenders may be sample variation, a small defocussing or any other influence. I would be confident the lens is fantastic anyways and the better IS system alone may compensate for this little slip up (if it's real)
- I don't think the is a possibility to judge on IQ based of bloc diagrams, the optical design optimizes so many parameter to get the optimum out f a optical concept, that discussing element sizes and assumed production tolerances of them it just unqualified guessing.
- the EF/RF discussion and assuming DSLR's are dead etc, are in the wrong place. The shooting costs when such equipment is used are so high, that a early replacement of a EF lens or having a special EF camera for this lens is peanuts in the overall costs, and the adapter is not that heavy and doesn't hurt when shooting a football game. So, I would not expect to see any ii versions at the next Olympics, Formula 1, or FIFA events. At such events, it's a small part of the cost to have the best equipment available. Even for amateurs, if I can afford shooting penguins in Antarctica, updating a camera or lens can not be the problem, I would go with the best equipment available, and do some training just before I go, to avoid handling errors.