In the Elsevier Boycott discussions, one thing which keeps coming up is: what value do the commercial academic publishers add? The general consensus is “not much”. They do some copy-editing of debatable value (the actual typesetting is now done by the contributing scientists for free), they do some administrative work… none of this should justify the four-digit costs they claim each individual paper incurs. Being the constructive type, I got to thinking: how can commercial journals add value in the twenty-first century? Here are the ideas I came up with. I’d be shocked to death if any of these were actually implemented, since that would subtract from the “all profit all the time” model which these publishers follow. But maybe these ideas could have some influence on emerging twenty-first century competitors.
Marketing
If we look across the aisle toward book publishers, we see they add a tremendous value to what they print: they market it. Compare this with academic journals. These offer nothing in the way of marketing. The Elsevier outcry has focused on the fact that the community does the writing, the typesetting, the reviewing, the editing, and the judging all for free. One more task which the community does for free is the marketing. Who takes the paper on the road and presents it at conferences? The author. Who cites the paper? The author and other authors. Who hypes it up with blog posts and comments? The author, and if it’s a really fantastic paper, others in the community. Who issues press releases if the paper is groundbreaking? The author’s institution and the community.
I’ve felt hesitant about writing this. The reason I felt hesitant is because we cannot even imagine academic publishers marketing for us without screwing it up (we shudder to imagine Elsevier spamming math bloggers with form letters). But as I thought about it, I realized, that isn’t how it ought to be. In a perfect world a highly-paid publisher ought to be qualified to talk about what they publish. If an Elsevier journal publishes a paper about self-distributivity and braid ordering, they ought to have in their employ some scientists competent enough to understand the paper. Of course, that might be expensive, requiring the publisher to house entire academic departments within its walls. But we’re talking about a giant corporation here, and how that corporation could earn its (very expensive) keep. If we can’t even imagine handing marketing to the publisher for fear of them screwing it up, then why on Earth do we let them publish us in the first place?
Proof Verification
The technology is still learning to walk, but one way a publisher could add very cutting-edge value to the math community would be to hire a bunch of technicians in Coq and Isabelle and other automated proof software, and formally verify accepted math papers. This is something of a pipe dream, but if the publishers pulled it off, then nobody could deny the tremendous added value. I recently taught myself Coq when a referee suggested I partially mechanize one of my own papers. It’s a fun language. I bet a lot of math papers could be formally verified by technicians with master’s degree levels of training in mathematics; if a paper is unverifiable because of too much handwaving, then merely alerting the author to this fact would itself be value-adding. Inevitably, the verification process would catch subtle glitches that eluded even official peer review, and that too would add terrific value.
To facilitate this, authors could submit two different manuscripts to the publisher. One would be the “human” manuscript where certain proofs can be “Straightforward.” or “Left to the reader.” or “By induction on n.” The other would be a larger manuscript where those omissions are spelled out, and this, while not itself being published, would be used by the Coq/Isabelle technicians. This would require more work on the writer’s behalf, but on the other hand… … … you guys ARE already checking those details on your own before submitting them as “straightforward”, right?!? (Guilty glances shoot around the room)
Translation
Here’s yet another way the big publishers could add tons of value: translate the journals into all different languages. Again, this would require they employ competent scientists, but remember: with great profit comes great responsibility. I, for one, would be thrilled half to death to have my papers translated into Mandarin and French and German and Japanese and Hungarian and… Hey, you know how Elsevier always says, “Acceptance of the copyright agreement will ensure the widest possible dissemination of information”? (cough cough BS cough cough) That might actually mean something if they had a department of competent bilingual technicians paid to translate your paper into exotic foreign tongues!
Analytics
“Translation? Lame, that’s so twentieth century.” Alright, back to the future, here’s what else the publishers could provide: advanced analytics like the ones Google offers webmasters. What might those analytics look like? Here’s a very rudimentary idea…
Analytics for your paper, “Counterexample to the ‘Circle is Round’ Conjecture”
- Published in: Transactions of the Kurdish Mathematical Society
- Traffic Sources (click for more detailed breakdown):
- MathSciNet: 27%.
- Google Scholar: 13%.
- Xamuel.com Eighty-fifth Linkfest: 11%.
- Other Sources: 49%.
- How Far Readers Got (click for more detailed breakdown):
- Abstract only: 43%.
- Abstract and Introduction: 25%.
- Up to first major theorem: 19%.
- Entire paper (not including appendices): 6%.
- Entire paper (including appendices): 7%.
- What Did Readers Spend Their Time On (click for more detailed breakdown):
- Abstract: 1 minute 23 seconds on average
- Introduction: 6 minutes 10 seconds on average
- Lemma 1: 12 minutes 53 seconds on average
- Proposition 2: 26 minutes 4 seconds on average
- Main Theorem: 42 minutes 12 seconds on average
Have any other ideas how the big publishers could justify their continued existence? Post ‘em in the comments!