I think in Harper's case he really does believe SML does it The Right Way, rather than promoting SML for ulterior/career reasons. His opinion on that subject isn't shared by many, but I think it's a genuinely held opinion. He just has very particular views on how type systems should be, and tends to overstate his views (e.g. treating questions reasonable people can disagree over as if he were stating unarguable facts).