Sometimes science is not enough...
It is clear to me that, high above most things, sharpness of a lens is a high priority in DXOMark tests. However, the sharpest of lenses do not always provide the most pleasing aesthetics/ look/ warmth/ bokeh in a picture. I have rented a Nikon 58mm f/1.4G and, whilst I am not prepared to pay the price for it (maybe if it was an f/1.2 I would), I can see what Nikon has done with this lens. It is worth remembering that this lens was in R&D for quite some time, and it was specifically designed to provide excellence in terms of coma control (in spirit of the Noct), chromatic aberration performance, minimal vignetting, ultra sharp stopped down (in fact, I don't know many sharper rendering lenses than this at f/8, for instance), and gorgeous bokeh. This lens does have all that in droves, and this lens also does produce a unique look to photographs that few others do... Subjects pop out at f/1.4... You don't have razor sharp sharpness of the subject at pixel level, but, you don't always want that (some people, for instance, don't like every line on their face sticking out like a sore thumb!). But, even at f/1.4 when you get the focus right, it is sharp "enough". The bokeh is absolutely stunning in this lens, really is, and as an astro/ night photography lens, it is also brilliant for the coma control. So, whilst DXOMark tests certain benchmarks, they do not provide the whole story.. You really do have to give a lens a test run first before you decide. Is it worth the money... no, in my view, but then what do I know - maybe the optics inside are expensive to make even if the outer body isn't, and I do appreciate their keeping the lens to a reasonably light weight too