That's just the thing: We don't see these as edge cases. We develop and maintain several commercial and private LOB applications that have need for more powerful tools than what WinRT can provide us access to. (Using a reporting tool is not an edge case in any business application.) Simply using an x86 tablet does not address the core issue.
The original complain from this part of the conversation was that the LOB talks at Build indicated WinRT was good for 98% of LOB apps, with 2% remaining on the desktop as legacy. I'm just saying that in our case, and in the cases of many other software companies we work with, the opposite is true. 98% remain on the desktop and 2% might be candidates for WinRT.
I just can't understand why the bridge to WinRT is so incomplete at this point. This is 2012 and we should be beyond what I would consider dramatic regressions.