There is a lot of confusion and inconsistency with citation of URIs. Sometimes, URIs are just hyperlink, sometimes footnotes, sometimes in the reference list. It should be possible to use URIs as “proper” references and for this you need metadata. This is why we created greycite.
Greycite (http://greycite.knowledgeblog.org/) is a tool which supports formal citation of content published on the web. It does this by attempting to extract the bibliographic metadata from the web page, and then storing it. Users of Greycite can retrieve this metadata in a variety of different formats. Further Greycite documentation can be found at Process (http://process.knowledgeblog.org/).
Greycite was developed by Lindsay Marshall (http://catless.ncl.ac.uk/Lindsay/) and Phillip Lord (http://homepages.cs.ncl.ac.uk/phillip.lord/) from the School of Computing Science, Newcastle University.
All metadata comes from the original URI. We believe this makes the metadata authoritative, in the sense it is sourced from the original author (http://www.russet.org.uk/blog/2012/05/greycite-citing-the-web/).
We support metadata in a number of formats. The easiest way, and the way that we have most extensively tested, is through the use of our own kblog-metadata plugin (http://www.russet.org.uk/blog/2012/03/kblog-metadata/).
Currently, we produce metadata in two formats. The JSON format is defined by Citeproc-js (http://bitbucket.org/fbennett/citeproc-js), and is used directly by the kcite plugin (http://knowledgeblog.org/kcite-plugin) that you can see active on this page.
The second form is bibtex (http://www.bibtex.org/), which is used by many tools.