This was also noted in 'The Debrief'. A colleague of mine sent me this article. I thought it was some sort of prank or 'meta' research to see precisely what some engineering and information science folks could get away with. In the case of the journal referees - apparently a lot. At first, the intro does seem to have a 'more-than-usual' issue with English grammar - but it 'improves' within the delivery of the mathematical arguments (or what simply measures up to affectations of them). Upon review, the maths statements make mostly no sense at all (I could see this generated by an LM). You have differentials that are operating on absolutely nothing, a single symbol epsilon that's takes on THREE different meanings, strings of mathematical arguments that are disconnected on so many levela - I mean the list goes on.

Sure, we see articles published that meet the general 'sniff' tests of referees, but I'm becoming completely despondent with the entire peer-review system. It's not like referees are paid a ton for their time - and that's the incentive to do so (and there aren't many who want to take up their time to do it). There are other 'issues' in the pub process, but moreover, there are mistakes that investigators make even unintentionally, but are enough to warrant retraction of the paper. We all must own our work. But I suppose I'm being too idealistic.

Given the current publication environment (don't think that journals like PNAS are immune either), things are going to fall through the cracks. However in this case, there was a gaping chasm in the process.