An article here, written by Dean Calver at the end of 2007, basically talks about the limitations of ray tracing. The good thing about this article is that it has been written from a neutral, unbiased perspective, which I find very rare.
As this article was written back in 2007-2008, it makes sense to evaluate it again, since things may have changed since then, so some arguments may no more be applicable and some new arguments may step in.
I would love to do that soon and post my views here.
In case the above link doesn't work, use this manual link: http://www.beyond3d.com/content/articles/94/1