Is there any point in a library paying for Advances in Mathematics?

Advances in Mathematics is one of Elsevier's top mathematical journals. If you are not at an institution that subscribes to ScienceDirect (Elsevier's portal -- if that's the right word), then you can buy a pdf from Elsevier for $41.95. This is a ridiculous price of course -- all the more so if the article is available in preprint form on the arXiv and shows up at the top of a Google search. Here is a summary of how many of its most recently published articles do not show up in some repository in one of the first few hits when you Google the title. I'll use the term "hidden" as a concise description of such articles.

Volume 253 (in progress, March 2014)

Three articles so far, one hidden (Dual systems of algebraic iterated function systems, by Hui Rao, Zhi-Ying Wen and Ya-Min Yang).

Volume 252 (February 2014)

Thirty articles. Four hidden.

I won't give all the titles of the hidden articles, but here are the authors.

Zhongwei Shen and Liang Song
Pavel Shvartsman

Jonathan Block and Aaron Smith (but there's a preprint of the same name that looks as though it is similar but a bit different and was submitted to Transactions of the AMS -- it looks as though it was rejected, then rewritten and successfully submitted to Advances)

Sabine Burgdorf, Ken Dykema, Igor Klep, Markus Schweighofer (this article's title starts "Addendum to ..." so is maybe not freestanding in some sense).

So that's five hidden articles (of which two could be partially discounted) out of thirty-three. Assuming that that is representative, there is a strong case for libraries to be told that we do not need them to subscribe to Advances in Mathematics. If they didn't, then the few authors who have not made their articles available would have a strong incentive to do so. 

In general, I've noticed that the better the journal, the more likely its papers are to be freely available online. So I would want to go further and say to libraries that we don't need subscriptions to mathematics journals any more. There are two big problems with that. One is that many learned societies depend on excessive and unnecessary library subscriptions to keep going. That problem could be dealt with in the short term if libraries just cancelled their subscriptions to journals published by the big commercial publishers. Then we could try to work out a more sensible way of funding learned societies.

The bigger problem is bundling. Even if all libraries in the world agreed that they could do without Elsevier mathematics journals, Elsevier does not offer a cheaper maths-free bundle. So the only way I can see a protest ultimately changing anything would be if mathematics departments made a huge fuss to their university authorities about being forced to contribute to the huge cost of journals that they don't need. But a great first step would be explicit declarations by departments round the world that it genuinely is the case that they do not need these journal subscriptions. I suspect that many people in decision-making positions don't realize this.
Shared publiclyView activity