The xiph.org article concentrates only on whether a given signal can be described sufficiently by 16/44.1. Several subtle and crucial differences between storage and playback of such a file have been pointed out in a white paper by Marco at HiFace and elsewhere.
The discrepancy between viewpoints is at least partly due to the variable behaviour of 16/44.1, 24/96, 24/192 and DSD files in each DAC. The bottom line is that is depends largely on which converter you're using and how it handles not only higher bit rates, but also bit rates based on 44.1MHz and 48MHz families (ie 24/88.2 vs 24/96).
So, yes, you could argue the difference is due to the format itself, but it would be truer to say that the real issue is how each converter handles the format, and less about how 'much information' it contains.
The discrepancy between viewpoints is at least partly due to the variable behaviour of 16/44.1, 24/96, 24/192 and DSD files in each DAC. The bottom line is that is depends largely on which converter you're using and how it handles not only higher bit rates, but also bit rates based on 44.1MHz and 48MHz families (ie 24/88.2 vs 24/96).
So, yes, you could argue the difference is due to the format itself, but it would be truer to say that the real issue is how each converter handles the format, and less about how 'much information' it contains.