refme

One of the necessary evils of academic writing are references: you have to add them, format them, keep them up to date, etc. Now, granted, the advent of LaTeX/BibTeX has made life a lot easier in this respect: you add a reference to your bib-file once, and from there on out you always have it at your fingertips in whatever referencing style you need. But ... you still have to add the references in the first place. I know, I know, this is quite the first world problem, but like so many first world problems these days, there's an app for that! Enter RefME, an app that promises to be "the free tool to generate citations, reference lists and bibliographies".

What does it do?

In a nutshell, it creates references for you. All you have to do is scan a barcode (of a book), enter a DOI-number, or simply search based on author or title and poof! a reference magically appears, in one of over 6,000 referencing styles for you to choose from. There's an iOS-app and a web-app, and once you have the references you need, you can export them to a variety of places like your clipboard, Evernote, Word, Endnote, BibTex, etc. It sounded pretty good, so I had to put it to the test.

Does it work?

I first decided to scan the barcodes of all the books I had lying about in my immediate vicinity:

  1. R. Zanuttini & L. Horn Micro-syntactic variation in North American English
  2. D. Geeraerts Theories of lexical semantics
  3. H. Smessaert Basisbegrippen morfologie
  4. C. Wheelan Naked statistics
  5. R. Munroe Wat als
  6. L. Gonick & W. Smith The cartoon guide to statistics
  7. F. du Bois & I. Boons Gin & Tonic: de complete gids voor de perfecte mix

I used the iOS-app to scan the barcode on the Geeraerts-book, and in the blink of an eye, the following appeared on screen:

@book{Geeraerts_2009, title={Theories of Lexical Semantics}, ISBN={9780198700319}, publisher={Oxford University Press, USA}, author={Geeraerts, Dirk}, year={2009}, month={Jan}}

Ok, not perfect—publisher's location is missing—but pretty good nonetheless. The same was true for numbers 1, 4, and 6: here and there information was missing, but it got the basics right. The app struggled, though, with the Dutch books on my list. In one case (number 5), it filled in author and title but nothing more, for another (number 3) it could only give me the ISBN-number, while the third one (the G&T-book) it didn't recognize at all. Then I entered some DOI-numbers in the web-app. This produced great results: all of them yielded perfectly formatted and complete references. Finally, I looked for some papers based on their authors and (keywords from) titles. It worked well with journal articles (I looked for Merchant's 2013 LI-article on voice and ellipsis and Adger's 2006 paper on combinatorial variability in the Journal of Linguistics), but behaved strangely with book chapters. I searched for "luigi rizzi fine structure left periphery" in the section Book chapters and found the relevant Haegeman-volume, but without title and author of the chapter filled in, while in the section Journal articles I found title and author, but it looked like the paper had appeared in the 'journal' Kluwer International Handbooks of Linguistics. A similar fate befell Kratzer's seminal 1996-paper on severing the external argument from the verb.

In short, mixed results: when it works, it works quite well and is highly useful, but there are also still clear gaps in RefME's bibliographical knowledge, some not unsurprising (I wouldn't expect a Belgian Dutch guide on how to mix the perfect Gin & Tonic to be in their database), others quite mysterious (like the Rizzi-paper).

How much does it cost?

Nothing. Nada. Bupkis. Typically, this is a cause for concern. Here's what RefME themselves say on their support page in reply to the question "How does RefME make money?":

RefME is very lucky to be supported by investors focused on growth. Our goal is to reach 10 million users within the next 18 months and we are already well on the way to reaching that number. We do know how to make money but don't worry, we aren't selling anyones [sic] data or working with any publishers (and never will!). RefME will also always be free to students :)

Translation: you'll have to accept on good faith that we'll be good. I actually wouldn't mind paying for this app, because it clearly has the potential of being very useful—imagine having to add 100 references to your bib-file in one go; RefME could speed up that process considerably—but then it has to become much more accurate.

  1. Don't judge me.

  2. The HPSG-community has quite a good solution to this conundrum: they host a central bibliography.