So what’s the Digital Object Identifier for, really? I thought it was a permanent identifier so that we could link from one article to the articles it references in a pretty seamless fashion. OK, not totally seamlessly, since a DOI is not a URI, but all we have to do is stick http://dx.doi.org/ on the front of a DOI, and we’re there. So we should end up with an almost seamless worldwide web of knowledge (not Web of Knowledgetm, that’s someone’s proprietary business product).
Obviously the Publishers must play a large part in making this happen. They support the DOI system through their membership of Crossref, and supplying the metadata to make it work. And sometimes they remember that when they transfer a journal from one publisher or location to another, they can fix the resulting mess simply by changing the redirect inside the DOI system. (And sometimes they forget, but that’s another story.)
And of course, these big, toll-access, subscription-based Publishers trumpet all the Added Value that their publishing processes put onto the articles that we write and give to them (and referee for them, and persuade our libraries to buy for them, and…). So obviously that Added Value will extend to ensuring that all references have DOIs where available? A pretty simple thing to add in the copy-editing stage, I would have thought.
Except that they don’t. They display few if any DOIs in their reference lists of “their” articles. In fact my limited, non-scientific evidence-collecting suggests to me that they probably do the opposite to Adding Value: remove DOIs from manuscripts submitted to them. OK, I have no direct evidence of the removal claim, but I reckon there is pretty good circumstantial evidence.
I don’t have a substantial base of articles to work from (not being affiliated with a big library any more), but I’ve had a scan at the reference section of several recent articles from a selection of publishers. What do I see?
Take for example this editorial in Nature Materials:
Nature. (2013). Beware the impact factor. Nature materials, 12(2), 89. doi:10.1038/nmat3566
Yes, there’s a DOI in the reference I used. Mendeley picked that DOI up automatically from the paper. If I use that paper in a reference, the DOI will be included by Mendeley. This presumably also happens with EndNote and other reference managers. (Here’s me inserting a citation for (Shotton, Portwin, Klyne, & Miles, 2009) from EndNote… yes, there it is, down the bottom with a big fat DOI in it.) (This is part of my circumstantial evidence for Value Reduction by Publishers! We give them DOIs, they take them away.)
Anyway, looking at that Nature editorial, there are no DOIs in the reference list. Reference 7 is:
7. Campanario, J. M. J. Am. Soc. Inf. Sci. Technol. 62, 230–235 (2011).
I tried copy/pasting that into Google. I get two results, neither of which appears to be a JASIST article. OK let’s try this one, in a completely different field, from an Elsevier journal:
McCabe, M. J., Snyder, C. M., & Fagin, A. (2013). Open Access versus Traditional Journal Pricing: Using a Simple “Platform Market” Model to Understand Which Will Win (and Which Should). The Journal of Academic Librarianship, 39(1), 11–19. doi:10.1016/j.acalib.2012.11.035
Again, none of the referenced articles have DOIs included in the reference list. Here’s a recent reference:
Jeon, D. -S.,&Rochet, J. -C. (2010). The pricing of academic journals: A two-sided market perspective. American Economic Journal: Microeconomics, 2, 222–255.
Maybe that article (and all of the others) doesn’t have a DOI? Same trick with Google, we don’t get there straight away, we get to another search, for articles with the word “perspective” in that journal… which does get us to the right place. And yes, the article does have a DOI (10.1257/mic.2.2.222). Let’s try this article; surely Nucleic Acids Research is one of the good guys?
Fernández-Suárez, X. M., & Galperin, M. Y. (2013). The 2013 Nucleic Acids Research Database Issue and the online Molecular Biology Database Collection. Nucleic acids research, 41(D1), D1–7. doi:10.1093/nar/gks1297
No DOIs in the reference list. Here’s an odd one, from Nature again:
Piwowar, H. A. (2013). Value all research products. Nature, 493(7431), 159. doi:10.1038/493159a
Here they include no DOIs for actual articles, but there are URL-DOIs for Figshare! The first two references are:
1. Priem, J., Costello, K. & Dzuba, T. Figshare http://dx.doi.org/10.6084/m9.figshare.104629 (2012).
2. Fausto, S. et al. PLoS ONE 7, e50109 (2012).
Do the latest OA publishers do any better? Sadly, IJDC appears not to show DOIs in references. I couldn’t see any in references in the most recent PLoS one article I looked at (Grieneisen and Zhang, 2012). Nor Carroll (2011) in PLoS Biology. But yes, definitely some DOIs in references in Lister, Datta et al (2010) in PLoS Computational Biology.
What about the newest kid on the block? You know, the cheap publisher who’s going to lead to the downfall of the scholarly world as we know it? Yes! The wonderful article by Taylor and Wedel (2013) in PeerJ has references liberally bestowed with DOIs!
When I tweeted my outrage about this situation, someone suggested it’s just the publishers simply following the style guides. WTF?
Publishers! You want us to believe you are adding value to our srticles? Then use the Digital Object Identifier system. Keep the DOIs we give you, and add the DOIs we don’t!
PS At one stage in preparing for this post I tried copying reference lists from PDFs and pasting them into Word. You should try it some time. It’s an absolute disaster, in many cases! Which is NOT the fault of PDF, it is the fault of the system used to create the PDF… ie the Publisher’s system. Added Value again?
PPS: here’s that reference inserted by EndNote:
Shotton, D., Portwin, K., Klyne, G., & Miles, A. (2009). Adventures in Semantic Publishing: Exemplar Semantic Enhancements of a Research Article. PLoS Comput Biol, 5(4), e1000361. http://dx.doi.org/10.1371%2Fjournal.pcbi.1000361
EDIT: As the comments below suggest, my post is generally true insofar as PDF versions of articles are concerned, although even there some publishers (eg BioMedCentral) do incorporate a hidden clickable link behind the reference (in BMC’s case to PubMed rather than the DOI). Several publishers have MUCH better behaviours in their HTML versions, with both explicitly visible DOIs and clickable versions of references). Sadly, HTML has no agreed container format, and is next to useless for storing articles for later reference, so it is most likely that the articles you store and use on your computer will be the sort of stunted PDFs I describe here. I still claim: this is not good enough.