The only question is when does it happen and who will lose out because of it.
The really important question is what would have happened if there wasn't a coordinated effort to help the "correct" fork?
It is important to keep in mind that this kind of problems are rare and have been already been thought about. Bitcoin as whole is resilient against this kind of problems. When two versions of Bitcoin nodes cannot agree on interpretation of the rules then miners vote with their feet. They either upgrade or downgrade.
The only thing interesting about this incident is that a small group of most potent miners have changed the outcome by dumping their version. It seems that at the beginning version 0.8 was winning and everyone would be forced to upgrade to that version on very short notice. Only when operators of big pools decided to downgrade, branch on which version 0.7 worked took over.
The issue got everyone by surprise mainly because old version (0.7) had a surplus rule/feature limiting block size. Was problem known before, developers would first release a version which accepts large blocks but does not produce them. Only when almost everyone would upgrade to that version, large blocks would be produced.
Actually this is what we are having now. Version 0.8 with default configuration accepts large blocks but does not produce them.