A sideline to the Internet protest against the ill-drafted US anti-piracy legislation has been that mathematicians have been spurred into action against big publishers. I am not certain of the history, but Tim Gowers seems to have taken the lead in a campaign targeted specifically against Elsevier, supported by Terry Tao and others. There is a website thecostofknowledge.com where you can choose to support the campaign in any of three ways, by refusing to submit papers, referee papers, or do editorial work, for Elsevier. The publisher is sufficiently concerned that a representative has written a long reply on Gowers’ blog.
I will return to this, but I am going to take a longish detour first. If you want to go straight to the website and register your support, fine.
The person who has had the greatest impact on mathematical publishing in the recent past is Donald Knuth. Trained as a mathematician (he wrote his thesis on projective planes under Marshall Hall), he became a computer scientist, where he may be best known for his magnum opus, The Art of Computer Programming. Unsatisfied with the typesetting of the books, he took a short time out to create his own typesetting system. The result is TeX, freely available to the worldwide mathematical community and, after not much more than thirty years, universally used by mathematicians.
This was an extraordinary achievement. As well as the typesetting program TeX, Knuth produced a font design program METAFONT, and the Computer Modern font family, designed so that the mathematical characters and symbols would go harmoniously with the body text. Furthermore, in the interest of encouraging good documentation, he produced his “Web” system, with two accompanying programs Tangle and Weave (I do not remember now which was which), one of which turned the file into Pascal source code, and the other produced a human-readable version incorporating the comments. The TeX and METAFONT programs were each published as a book to illustrate the idea that even a very large program could be structured and documented using these tools.
The manual for the program, The TeXbook, evokes mixed reactions, but in my opinion it stands head and shoulders above any other software manual I know. It goes from the most elementary to the most arcane aspects of the program, and is written with style and charm. It includes self-reference: each chapter ends with a couple of apposite quotes, and one of these is taken from the book itself. In addition, the introduction promises that the book contains both jokes and lies. (Jokes and lies? In a software manual?) The very last exercise in the book reads
EXERCISE 27.5 Final exercise: Find all of the lies in this manual, and all of the jokes.
The first appendix is entitled “Answers to All the Exercises”. Turning to the appropriate place, we find
27.5. If this exercise isn’t just a joke, the title of this appendix is a lie.
The reason I have diverted to discuss this is hidden away at the start of Chapter 16, “Typing Math Formulas”:
Notice that all mathematical formulas are enclosed in special math brackets; we are using $ as the math bracket in this manual … because mathematics is supposedly expensive.
Indeed, in the days of hot metal typesetting, mathematical typesetting was a specialist task for which printers could charge premium rates. Perhaps Knuth’s greatest achievement was to give us the tool to do our own mathematical typesetting, with no charge at all!
TeX, in the particular format of LaTeX, eventually became the universal default for mathematical typesetting. Not only did mathematicians take to the convenience of the TeX conventions for mathematics in emailing mathematical content to one another, but even the largest publishers produced style files and accepted submissions in LaTeX (often exclusively so).
As well as not having to provide expert mathematical typesetting, publishers now (as a result of the internet) have less need to provide printing, warehousing, and distribution. However, their charges to the mathematical community for their services do not seem to have been reduced to reflect this, and their marketing practices aim at maximising their revenue at the expense of wide dissemination of information. It is because Elsevier are felt to be the worst in this respect that they have attracted the protest. (A Dutch colleague once told me, “All Dutch publishers are crooks”; since then the other large Dutch academic publisher, Kluwer, has become part of the Springer empire.)
The mathematical protest seems not to have created many ripples in the wider world yet, but the internet protest against ill-judged lawmaking on piracy has done so; the Wikipedia blackout affected many people. Nature last week devoted their first editorial to the affair. They are trying to position themselves as on the side of the angels. They write,
No one disagrees that a publisher of review articles deserves to charge for access to them. After all, the publisher’s staff have contributed value in various ways, identifying the author and the article’s aim, assessing and editing the draft, selecting peer reviewers, working with the author to build on their advice, developing illustrations, rendering the article into print and online forms, maintaining it online and including links, citation statistics and other enhancements.
I can hear the hollow laughter from mathematicians reading this. I consider my review articles to be probably my most important contributions to mathematics. No publisher has ever approached me for a review article; it has often been a hard task for me to convince the publisher that they want to publish it. Choice of referees is done by the (unpaid academic) editor, not the publisher’s staff; and the referees (who do the real work) are also unpaid academics. Again, the editor mediates between author and referees, and the publisher is presented with the finished product ready to be put on the website. The article usually does not appear on the website for some time because the journal has a backlog. And finally, I think anyone reading this will probably know my views about citation statistics and the damage they cause.
So to the boycott. It sounds fine in principle, but what will it achieve? Said otherwise, what event would be regarded as a successful conclusion? A few mathematicians are hardly likely to bring Reed Elsevier to their knees. Also it is not entirely clear what freedom authors have. The Elsevier journal Discrete Mathematics has published selected papers from the British Combinatorial Conferences for many years. Such publication is valuable to young authors, but it does require a lot of work by the guest editors and quite a bit of goodwill on both sides. Moreover, there is no viable alternative place for such publication. If the protest is directed against the bundling of journals, the publisher can easily offer individual journals but increase the price to keep up their profits; the impact factors of some journals might take a hit, but the popular ones will be unaffected.
Publishers say, “Publish with us; look at our impact factor.” Mathematicians say, “I don’t care about impact factors,” but increasingly the bureaucrats say, “Oh yes you do.” We were recently required to produce a list of “aspirational journals” to which we would submit our papers. Our research director, quite correctly (in my view) wanting to deal with this quickly so as not to eat into his research time, took the Australian Research Council listings (which the Australians have now disowned, by the way), and invited us to suggest additions to the A* category to use as the basis of our list. People of the stature of Tim Gowers and Terry Tao, and people of my age, can tell the bureaucrats where to put these lists; but not all our colleagues have this option.
The Nature article recognises two viable modes of running journals: subscription, or author-pays open access. Pressure from national bodies in some areas is likely to push us towards the second. This will produce a two-tier academic system, where academics at rich universities, or those with grants including money for page charges, will be able to publish, and others will not. However, there are two further modes: free and unrefereed repositories like the arXiv or personal web pages, and free refereed journals run by volunteers. The first of these two is officially supported by the research councils and some universities in the UK. (Our institutional publications list takes a feed from the arXiv.) When I raised these issues in the very early days of this blog, Laci Babai eloquently and passionately defended the last solution, journals run by volunteers; his own journal, Theory of Computing, is a fine example.