M Evening says filters on UWA degrade IQ

I've used a B+W UV filter on my 10-22 (as I do on all my lenses) without giving it a second thought. Reading M Evening's recent book on the CS4 Photoshop, I found he writes that filters cause IQ degradation on UWA lenses, particularly on the edges. This is not about vignetting. The argument is that light coming at a low angle needs to first traverse a good part of the glass filter before hitting the front element. I'd like to know your opinion on this. Have you tested your UWA lens with and without filter and noticed something remarkable about the IQ on the edges? Thanks in advance.

And what are the results of your tests?

UWA zooms are all soft in the corners.

I think I'd worry about it with a prime where I wasn't going to crop it off anyway.

Degrades - maybe - changes what gets to sensor - probably .
There is nothing to replace my CPL .
Digital sensor are not supposed to see UV - I still use them - they seem to cut down on glare , flare and reflections and it's easier and faster to change a filter than clean a lens .
It's next to impossible to take same picture with and without filter - even in a studio , the light changes because the electrictricy is changing 120 times a second .
It comes down to you deciding - I use B+W on all my lenses -