I maintain that this is a grey area that has not been tested in courts, mainly because ACM would not pursue something like this. In a copyright lawsuit, it is n

I maintain that this is a grey area that has not been tested in courts, mainly because ACM would not pursue something like this. In a copyright lawsuit, it is not obvious to me that a judge would agree with your conclusion that slight differences in formatting make it "not the published version". INAL, but to me, the mention of ACM on the paper is itself problematic from a copyright standpoint. Of course, the larger point I'm making in all this is that this is completely unnecessary and regressive.