I claim a red sheet metal screw. Suppose there are no red sheet metal screws. Suppose there's no evidence that any ordinary artisan at the pertinent time had appreciated any reason why red sheet metal screws would be advantageous.
Sheet metal screws are known. Red paint is known. It is known by the ordinary artisan that application of red paint to a sheet metal screw in a conventional, known manner would result in a red sheet metal screw.
Hence, red sheet metal screws are obvious.
Maybe. But saying combinations of known elements are obvious without providing a benefit or reason to combine them just doesn't seem, well .... fair. To me, that's tantamount to saying that just because a POSITA *could* combine the elements means that he *would* combine the elements.
There are thousands of combinations of types of screws and colors of paint. Out of all those thousands of combinations the inventor chose a particular kind of screw -- sheet metal -- and a particular color -- red. Why is obvious to pick that one?
Now, if there was no particular reason that the inventor selected that combination, then maybe it's a "mere design choice," and maybe he doesn't deserve a patent.
I will comment on the part that I emphasized. It isn't obvious why one would want to paint the screw that particular
color indeed. But if the patent application doesn't help to answer the question "why not other color?", then it's a
predictable result. It's not that the choice to paint it red is obvious, but the application doesn't disclose anything
over what's known/obvious about such combination.
But if the inventor can show that he had a reason for picking that combination (probably characterized as a "benefit"), then doesn't he deserve a patent?
If the application discloses anything unpredictable about the combination or method of combining, then
I think at least this rationale cannot be used to deny the patent.
EDIT: You have brought up the novelty with relation to this rationale, and thinking of it in retrospect, I think
an interesting way to think of novelty and obviousness is both belonging to the same "difficulty" scale:
Well established --- Prior art --- Novel, but a predictable result
--- Novel, but obvious --- Novel, non-obvious (hello invention!)
The bolded predictable result rationale is on the border between novel and prior art, between 112 and 113.