Ah, you see we are approaching this from different directions. It's a bit like the ITIL thing on another thread... you're focussed on what light reflects off what bits of the environment and where about those bits of environment are relative to the lamp. I'm focussed on (an I believe the article's author is also referring to) the HUMAN aspect of lighting. I'm not saying that I couldn't see my friend standing 3 metres inside the park because the new LED lamp wasn't shining on her, I'm saying that I couldn't see my friend standing 3 metres inside the park because all the blue light between myself and her had screwed up my scotopic vision and buggered the dynamic range of my eyes so that the tiny amounts of light scattering and bouncing around the environment and reflecting back off her are pretty much wasted when they get into my retina.
The practical upshot of all of this is that it's all well and good saying "Well, for these 4km of M-class road between points A and B we meet x, y and z standard of illumination", when Joe Bloggs is driving along in his car from A to B but then leaves the M road and proceeds along another road from point B to point C which meets some other version of lighting standard x, some 300m later, where he turns off into an unlit unclassified road (which doesn't meet any illumination standard) heading towards point D and his bed and flattens Jane Doe because he didn't see her under the half-moonlight because his vision hadn't yet adapted back to a scotopic or mesopic range. The answer is not to then light all roads to standard x, y and z, but to go back and revisit what the effect is of standard x, y and z when humans move from one condition to another. If there's a fault, it lies in the standards and those who apply them unswervingly forgetting that there may be confounding factors at work.