We used to have Postgenomic.com, and we have the PLoS journal's article level metrics for comments, the pay-walled F1000, the free ResearchBlogging, and the option to leave notes and comments in bookmarking systems like CiteULike and Mendeley. And we had tools to visualize some of those on a paper's HTML page (see this paper). Some papers are retracted, and that news is not always clear either (there is RetractionWatch), and aggregation of that would be useful too.
But, what I am really looking for is a service that finds comments for a given paper, aggregating from various sites. I am aware of none. What is at this moment the best way to find all (public) comments made on a particular paper?
(See this question on where you can publicly post your comments on a paper.)
Be aware of blatant self advertisement. But we have built a service which will not allow you to find comments about your paper, but it gives you insight where you paper is cited outside the literature domain. http://www.citedin.org aggregates citation information from about 50 online resources (blogs, databases, wiki's)
There is no one tool that can do this, except perhaps for google. The main issue is that many comments on papers can occur outside of any of the sites where the paper appears, for instance on blogging platforms or self-hosted blogs, on news sites (think of the comment threads on a Guardian article talking about a paper).
Ideally the original publisher page for any article should accept trackbacks or updates when a comment appears about that article, and whether it displays those or not, it should provide a feed for others to consume. Hand in hand with this, sites that comment on that article should push their conversation back to the parent article. If that sounds a lot like PubSubHubBub + Salmon, then that's what it is, but that layer of infrastructure seems to be taking a little time to diffuse into the systems that we use for commenting on the literature, and in addition, it's is going to be longer before publishers put in place the kinds of endpoints that I describe here.
I've been having some fun with that, liking it. I also like the Altmetric that's an app in the Elsevier system. Here's what it showed for the BioStar paper: http://screencast.com/t/by4TjQ1JVhp8
For reference: Are There Websites Which Allow Users To Post Comments On Peer-Reviewed Articles?