A new vision for Mendeley
by Edward Z. Yang
I use Mendeley because it lets me easily search for papers I care about. Unfortunately, that seems to be all Mendeley has been doing for me... and that’s a damn shame. Maybe it’s because I’m an undergraduate, still dipping my toe into an ocean of academic research. Mendeley was aimed at practicing researchers, but not people like me, who are stilll aiming for breadth not depth. I can count on two hands the number of technical papers I’ve really dug into—I’m trying to figure out what it is exactly that I want to specialize in.
From this perspective, there are many, many things that Mendeley could be doing for me that it simply isn’t doing right now. And this is not purely a selfish perspective: the world of academics is a relatively small one, and I’d like to think a move towards my preferences is a move towards a larger body of potential users and customers. My request/plea/vision can be summed up as follows: Mendeley needs better metadata.
Metadata extraction has been a long standing complaint of many, many Mendeley users. One might at least partially chalk it up to the obsessive-compulsive nature of researchers: we want our metadata to be right, and having to manually fix essentially every paper we import our database creates a huge extra cost on using Mendeley.
How correct should this metadata be?
- If I search by author name, the list of papers I get should be equal in quality to that author’s listing of publications on his personal website. (Name de-duplication is pretty terrifying: Mendeley gives no indication of what permutation of a name is likely to be the right one and gives no advice about whether or not initials or full names should be preferred). As a prospective graduate student with some degree of specialization, knowing which authors I have the most papers of gives hints as to what departments I may be interested in.
- If I search by conference name, the list of papers I get should be equal in quality to the table of contents for that conference’s proceedings. Furthermore, there should be continuity over the years, so that I can search for all-time ICFP, not just 2011’s proceedings. Conferences have a notoriously large number of permutations of spellings (is it ICFP or International Conference for Functional Programming or...), and with no canonicalization keeping these straight is hopeless.
Here is something that would be really interesting to extract from papers, that is not readily available anywhere on the Internet: who funded the piece of work, and under what program! Researchers are obligated to acknowledge their funding source in a footnote or a section at the end of their paper, and access to this metadata would let me know “Who funds most of the research in this area” or “What grants should I be looking at for the next funding cycle.” This information is, of course, currently passed down as folklore from advisor to advisee. Extracting, polishing and publishing this data would be an interesting endeavor in its own right.
Mendeley does a great job letting you attach metadata to a specific paper, which is searchable but not much else. But there are plenty of small chunks of information that could be profitably converted into generalized schemes. For example, Mendeley currently stores a bit indicating whether or not a paper is “read” or not. This is woefully insufficient: one does not simply read an academic paper (nor does one simply walk into Mordor). Maybe you’ve read the abstract, maybe you’ve skimmed the paper, maybe the paper is one that you are actively attempting to understand as part of some of your research. This workflow should be made first-class and the relevant user interface for it exposed. This information is also important if you’re trying to draw inferences about your paper reading habits from Mendeley’s database: papers you’ve imported but never looked at probably shouldn’t be counted.
In many ways, this is the call for the equivalent of the “Semantic Web” for academic papers. By in large, we are still attempting to realize this vision for the Internet at large. But there are good reasons to believe that the world of academic papers may be different. For one thing, the web of academic papers has not even reached the point of the traditional Internet: it doesn’t have hyperlinks! Additionally, academic papers are much slower moving than traditional web content and far more permanent. I want there to be no reason for my friends in industry to complain that they’d much rather a researcher publish a blog post rather than a paper. It’s ambitious, I know. But that’s my vision for Mendeley.
Did you enjoy this post? Please subscribe to my feed!