As far as I remember the article said that it used to be that the greatest DR was to be had over at the right hand side and ETTR therefore was a real advantage. Their thrust was that sensor design has moved on and that's no longer the case.
I wish I could remember where I read it but all I can remember is that it was on one of the usual suspects sites.
Not that my cameras are cutting edge, they aren't, and regardless of any finite advantage or not I think I'll carry on until I see a disadvantage other than slowing the shutter speed. With a CSC it's easy as you can have an in view histogram and I can ETTR if possible and get a better first time keeper rate in difficult lighting with my G1 than I can get with my superior 5D.
Not the greatest dynamic range to the right of the histogram, that's a contradiction, but a load more data recorded on the right and that cannot change. Eg, if the dynamic range runs from 1-256, that's a difference of eight stops (1, 2, 4, 8, 16, 32, 64, 128, 256) from dark to bright. If you draw a line for mid-grey somewhere between 16 and 32, in round numbers there will be at least ten times more data recorded on the right.
My point, and perhaps the point being made in the article you refer to, is that by using ETTR and adding maybe two stops more exposure (often possible) the lower end values of 2 and 4 etc become 8 and 16 - quadrupling the amount of data captured. This makes a big and visible difference to signal-to-noise ratio in the shadows with attendant benefits, whereas packing yet more data to the right where there is already very low noise, has little or no visible advantage.
Perhaps ETTR would be better named PUSFTL - Pull Up Shadows From The Left

but as technology improves these upsides will become eroded.
Getting back on topic, the basic technique behind the Zone System, in very broadest principle, is to expose for the shadows and then hold back the highlights in development. ETTR works differently, but end the result is often similar.