Why the DxO bashing?
The bashing are performed by insecure brand owners.
Well, maybe. But I currently own Canon FF & APSC, Sony FF and Olympus Micro 4/3, and have previously owned Nikon and Pentax....
Others have explained better than I can the various biases (i.e. the criteria they use) of their evaluation process, but even leaving that aside, by trying to reduce all of this to a single number they're making the same fundamental mistake that everyone else makes when doing so (and not just in this context, of course), namely trying to reduce measurements of properties that are incommensurable to some point on a common scale and, in the process, omitting some properties altogether. For instance, the five qualities of a lens that they explicitly provide numbers for (sharpness, transmission, distortion, vignetting and chromatic abarrations) may each be measurable, but they're not measurements of the same sort of thing on the same scale, so you can't just add them up (you might as well try to add your weight to your height - it's a conceptual absurdity). And if you care about coma, say, or bokeh (try measuring *that*!), well, they don't seem to figure in at all.
To the extent other sites do this, their scores are absurd too, for the same reason. I'm not sure why DxO's has become the reference, though, often invoked by other reviewers, bloggers, etc. Maybe it's in part because, to an untrained, casual observer, the presentation looks so scientific - lots of charts and numbers and nothing as preposterous as testing a lens on a camera or a sensor in a lens/camera, or as vulgar as showing actual photos taken with any of the equipment reviewed (or am I missing something?). To state the obvious, you can only take a photo by putting a sensor in a camera and attaching a lens to it; how any one of these components "performs" (in some weird sense of the term) in isolation hardly matters. Which is why the most useful (to me, anyway) review sites provide photos (preferably comparative) to prove their points, and why ultimately there's no substitute for trying equipment first-hand. It doesn't matter if lens A has a better score somewhere than lens B if, for your purposes, you can't see a difference that matters.
Oh, and like many others, I think DxO's software is good (probably the best for noise reduction).