You are right justified in making this observation (it occurred to me too later last night).  Albie's find is valuable from the sandpoint of that exerpt appears to defeat one of the claims of prior art, but doesn't touch on the "operands must be left or right justified" issue (as opposed to our preferred interpretation of "optional"/either/or).  So ARM may not be perspiring quite yet.  Maybe there's something else in that package that hits this point.

Good eye!  And memory!

SGE