Assuming the above graph is relatively close to the truth, IE8 is around 10% of the worldwide marketshare.
So, the question is: when does a library like jQuery drop support? 10% marketshare, 5%, 2%, 1%? At what point is the overhead outweigh the benefits?
Obviously there are some developers who need to support IE7 (for example), regardless of the overall market. I personally don't think that should be the approach for tools like jQuery: time is valuable, and they should target the largest possible market possible.