Connotea is a Database
Or, put another way, Connotea is for much more than just bibliographic references.
As Jon Udell so ably demonstrated in his recent post (Del.icio.us is a database), the ability to tag items with freely chosen keywords, coupled with the ability to mix-and-match queries on those tags, gives you a powerful database that is good enough for a large range of purposes.
Given that Connotea has these capabilities (and more besides — see below), can Connotea be used as a database too? Absolutely. And happily, Connotea users have already realised this and started applying it to their own specialised needs.
For example, Rod Page wanted to be able to correlate TreeBASE taxon IDs to the original description of the (in this case fossil) species. His solution was elegant and simple — bookmark the original description in Connotea and use the taxon ID as one of the tags. This solution is also highly flexible: by using multiple tags and bookmarks, he could collect together descriptions and databases entries with other relevant items about that taxon to form a kind of browsable mini-database for that species (although I don't think he's explored this yet).
Meanwhile, Robert Muetzelfeldt, an ecological modelling researcher at the University of Edinburgh, is using Connotea as a database in a different way. He's constructing a database of ecology and structural biology models represented in XML by bookmarking links to those XML documents and using tags to label the format and content of the models. This approach allows him to easily collate and query information that would otherwise be highly distributed. Want to find a list of structural biology models involving cyclin? Easy, just have a look at http://www.connotea.org/user/robertm/tag/SBML-XML+cyclin. And using Connotea to create a database of distributed data doesn't stop there — in a lovely piece of self-reference, Robert's even using Connotea to point to stylesheets that let users create custom views on Connotea!
The great thing about Connotea is that the entire database is open to querying and manipulation via the Web API. This means that, in principle at least, many users could collaborate on these databases, and you could run complex queries across the entire collection. So even if you can't do some queries through combining tags (like negative queries, for example), you can still get the answer by writing a quick Perl, Python, Ruby, or any-other-language-you-like program to sift through the data. There's also the raw RDF output, which you can query using SPARQL, as Rod Page has demonstrated elsewhere.
One small, but significant, feature of Connotea that lends itself nicely to all this is the tag note. A tag note lets you explain to other users how you're using a tag, so as well as being able to quickly build an online database by bookmarking and tagging things in Connotea, you can also explain how to use it at the same time.
