2010 October 3
While getting ready for my qual, I started using the website CiteULike to keep track of papers to read. It’s one of many pieces of reference management software out there that are meant to help scholars keep track of the articles, books, websites, etc. to which they refer in the course of their work.
In my experience, it’s hugely valuable to think of programs in terms of the data they generate or modify, and from this perspective all reference management programs are extraordinarily simple: they maintain 1) a database of bibliographic information and possibly 2) some of the referenced content (i.e., a bunch of article PDFs). To me, data like these cry out to be stored in the cloud and accessed through a web app: there are all of the usual benefits of online storage, and the operations on the data are so straightforward that it’s really hard to imagine that a client application is actually necessary with 2010 (or probably even 2005) web browsers.
There are several features that I’d like to see in a web-based reference management tool that CiteULike doesn’t have, and as far as I can tell none of its peers do either. The most important is good integration with bibliography databases like the aforementioned ADS. In the summary page for every paper I’d really like to see a list of its references with links to their abstracts. I’d like to be able to run a search like, “Show me papers that are frequently cited by the papers already in my library.” Other tools out there seem to have better support for this, though I don’t get the impression that any of them have integration as tight as I’d like to see.
Another feature I’d like to see is automatic background retrieval and storage of article PDFs, so I can know that I can always get my hands on the article itself; this would also allow the reference manager to do full-text searching of the articles in my library which would be pretty nice. Most reference managers let you upload PDFs manually, but all of the information needed to fetch the PDFs directly is available through modern bibliographic metadata.
My interpretation is that most reference manager tools don’t support these things for the same reason: the bibliography databases and articles are copyrighted, so your app can’t generally just go ahead and download the information you’d like. (There are important exceptions: preprints on ArXiV are freely available, of course, and ADS is a free service, though I think there are rate limits for clients.) It seems to me that there’s a pretty obvious solution of having the user enter login credentials for bibliography and full-text sites if they have them so your app can at least optionally take advantage of the services, but for some reason few of the desktop apps and none of the web-based apps support this, as far as I know. (For app authors, it’d definitely be a hassle to have to integrate with a bunch of uncooperative services, but it seems like it’d be well worth the effort.)
One thing that the desktop apps do that the web apps don’t — yet — is have integrated reading of article PDFs. That kind of functionality is hard to replicate on the web, but sites like Google Books show us that you can put something together that’s at least competitive with desktop readers. I’d be interested to see a sincere effort to do really good web-based article- reading. I think a webapp that did a really good job of integrating article reading, citation management, and exploration of the bibliographic database would be really, really compelling.
Even without these wishlist items, I still find CiteULike to be a great tool — I’d say that reference management is one of those problem domains which maps really, really well onto the web app space. I wouldn’t say that it’s clearly superior to any of its peers, but it certainly does the job well. (One nice thing is that all of these tools support BibTeX import and export, so you don’t have to worry about lock-in.) I don’t think the pieces are in place quite yet for CiteULike or any of its peers to be must-use scholarly tools, but with some good design and development effort I can totally envision them becoming that good.