With regard to calculating link strength, I think one of the steps has been improved ability to do visual page segmentation - that is, "reasonable surfer" model rather than the "random surfer." The more a link is displayed in a way that a reasonable surfer might click on it, the higher the value transferred.
I also expect this area to be one of continued evolution for Google. For example, not just the greatest power for "in content" links, but "in content and above the fold."
Panda may not have been measuring "above the fold" directly (comments from Matt Cutts at PubCon seem to indicate that), but if they get the algo for it nailed down, I can imagine them being delighted to fold it in as a direct factor, rather than trying to measure its effects indirectly.