Difficult to deal with references in an article, they’re locked up, so that prevents some discovery.
Martin Fenner – need to have a semantic meaning for citations
Melody Dye – citations fall under a Zipfian curve. we like to think how we cite is logical, but it’s often the principle of least effort (a la Zipf).
CiTO – citation typing ontology
Q need citations in context, how do you do that? (part of CiTO does that)
A need also where in the text, there is a semantic web solution for that.
A2 swan project from Harvard – semantically marking up Alzheimer papers
Q: why can’t a citation say I disagree with this knowledge claim, not I disagree with this paper - needs to point to a specific section
Q: citations have this level of respect so they provide a grassroots way to endorse non-traditional products – some journals scream but others don’t.BUT when you do that, the aggregators rip those out (like WoS getting rid of dataset DOIs). What does Mendeley plan to do?
A: Mendeley hasn’t approached this yet. They took out the reference extraction part, but they are putting it back – but this won’t change anything that’s in the original document.
Q: WoS and Scopus aren’t transparent – about open bibliography – is there a way we can build this openly.
Q: open bibliography – does it really matter if it comes from a commercial entity
a: I need to be able to devise open algorithms for evaluation using the open bibliography.
a: the algorithms and the data should be separate
mendeley’s business model will be to free these semantically marked up data but charge for some of the analytics
q: in terms of sentiment – is there evidence that scientists will be as critical in something that goes somewhere. Sentiment always appears more positive than it really is.
a: (from me) have been attempts to do machine sentiment analysis
q: there’s no real comprehensive theory of citation (he talked through Nicolaisen’s review). lots of reasons to cite - how can we account for this in the ontology
q: about students – teaching them the meanings of citations
a: from Bonnie Swoger teach examples and talk about citation best practices
q: comment about hassle with number citations
a: can just have mouseover of the citation
q: maybe this current system is the best possible compromise. in the beginning of the web there was the whole idea of negative links and stuff and it fell out
a: technical easy. move to adding structure to links. google has said we can look at rdfa. if there’s a payoff, people will do the work.
there are less than 40 semantic markups that could be used, the swan project is looking at these.
q: why do we need 1000 citation styles
a: the humanities folks fight that idea. but even in science – who can make that change?