Or love me less, or love me more
And play not with my liberty
Either take all, or all restore
Bind me at least, or set me free
  —Sidney Godolphin

It's hardly a secret that commercial scientific publishers are adept at gouging universities and other institutions for big bags of cash. So perhaps we should simply be grateful when articles are available free of charge.

But there's free, and then there's free.

I've recently noticed some journals where all the content is free on one website, but a publisher is charging for access to the same stuff on another. Two examples: the legendary Organic Syntheses and, slightly off my personal beaten track, the Journal of Allergy and Clinical Immunology. Visit the Org. Synth. website, say because you want to whip up some CBS catalyst, and no matter whether you're accessing via diamond-encrusted WiFi from the tallest ivory tower of your storied academic institution, which subscribes to everything up to and including Heterocycles, or squinting through the gloom in a grubby internet café in Ulan Bator, you can see the paper. Try to access the same article on Wiley's site, on the other hand—no dice, unless you're a subscriber.

At this time of year hay-fever sufferers might well want to read up on fluticasone furoate, and J. Allergy Clin. Immunol. is as good a place as any. Once again, on the journal's own site, the paper is free, but over on ScienceDirect, Elsevier will try to skin you for $36. Note that we're not talking author manuscripts here; the free article is the typeset, final article of scientific record.1 Nor are we dealing with articles uploaded here and there, (technically) in breach of copyright. These are official sites hosting the entirety of the journals' content, with, I can only assume, the full knowledge and consent of the respective publishers.

Now, an obvious response would be that since both are there on the internet for people to see, the market will vote with its feet. The problem with that, though, is one of discoverability. You can only take advantage of the free version if you can find it. Neither of these examples are particularly hard to find… if you know to look for them. Very often, people will arrive at a paper via a DOI—if you come from SciFinder, Reaxys, Web of Science, PubMed or anything like that, you'll be directed to the paper through a DOI resolver. And guess what? The Wiley Org. Synth. articles have different DOIs to the free ones. The literature databases that I tried did contain the DOI pointing at the free Org. Synth. site, which is something, but I've no doubt that the Wiley DOIs are out there, lying in wait for the unsuspecting reader.

The J. Allergy Clin. Immunol. example is odder still. I could only find one DOI per article (10.1016/j.jaci.2007.02.022 for the article linked above), but it can lead you to either site. However you end up following that DOI (SciFinder etc., or directly from dx.doi.org), it sends you first to ‘linkinghub.elsevier.com’, which seems to make a decision where to redirect you based on what type of user you appear to be. At home (not quite the Ulan Bator internet café, but with a similar level of journal access), I was redirected to the site with the free articles. Hurrah! But at work, the same DOI sent me to ScienceDirect and their paywall. I assume this is because ScienceDirect saw me as a ‘customer’—we have IP-authenticated access to a handful of Elsevier journals, none of them J. Allergy Clin. Immunol.—but why that means we can just be ushered away from the free content is baffling. Two questions spring to mind: what would CrossRef make of the same DOI taking you to two different places, and more importantly, why does the ScienceDirect journal page exist at all? Selling something that you know to be freely available elsewhere is fundamentally dishonest.

I don't know how widespread this kind of set-up is, and I don't know why these journals have this weird arrangement. Org. Synth. is admittedly a bit of an oddity in other ways, but J. Allergy Clin. Immunol. is, as far as I can see, a fairly standard sort of journal. Whatever the reason for it, it seems that once again in scientific publishing, the customers—researchers, librarians, funders—are getting the short end of the stick.

1. The value of an edited article vs the authors' original is another question for another day.