I'd be more enthusiastic about the Viking approach -- make as few assumptions as possible, and use experiments that have some likelihood of telling you something interesting about soil chemistry even if no organsisms.

Would anyone who thinks life may be in the Martian near-surface really be convinced that it's _not_ there just because a rover didn't find identifable DNA in the soil?

It's not that I think testing for complex biomarkers like DNA or ATP is wrong in the abstract, but given limited experiment opportunities, the high probabilty of getting just an inconclusive no makes me very skeptical that I'd use the money and mass budget that way if I were in charge. But I'm not, so we'll see what they find. :)