A typical computer science paper might represent the work of 2-4 authors over a year. Even though these authors don’t spend 100% of that year working on the paper, just counting their salaries, benefits, etc.. we see that the total cost to produce a paper can still easily amount to several tens of thousands of dollars or more. This cost is typically borne by a combination of their employers (often non-profit or public universities) and funding agencies such as the NSF.
When the paper is submitted to a conference, referees and the program committees typically spend some additional few hours reading it, debating its merits, writing reviews etc.. They are not paid for the effort, and so this additional cost, amounting to several hundreds of dollars, is also typically borne by their employers. Finally, the paper is published, which in these electronic-proceedings days amounts to placing it on the web (perhaps after asking the author to covert it into an ugly two-column format) . This costs about $10 per paper on the arXiv, or about $400 for the ACM (see here). Thus the publication cost is roughly somewhere between 0.1% to 1% of the total cost to produce the paper.
There has been much debate about who should spend this 1% of costs, and whether spending it entitles organizations such as Elsevier or ACM to collect royalties from readers of the paper. In this debate, ACM’s positions are mostly not supportive of open access (e.g., see here and here). Indeed in some sense ACM’s policies are more restrictive than Elsevier’s, who allows free access to mathematics papers after 4 years from publication. ACM also opposed the Federal Research Public Access Act (FRPAA), which would have mandated open access to federally-funded research papers after 6 months of their publications. The ACM also initially refused to oppose (and even implicitly supported) the Research Works Act (RWA) which would have prohibited open-access mandate for federally funded research.
Since these days almost everyone puts their papers on electronic archives, perhaps these backward stances of ACM don’t have too much negative effect on actual access to papers. But they may have very negative effects on the ACM itself. First and foremost, by acting against the basic instincts of scientists, who want their works to be public and accessible, ACM is risking that conferences and perhaps eventually entire communities may just decide to leave. ACM is also risking its financial future by refusing to acknowledge that, whether it wants it or not, open access is the future, given that more authors are putting their papers on their web and more institutions and governments are considering open access mandates. But, at least to my knowledge, ACM is not taking the necessary cost-cutting steps to reduce their cost structures to levels more similar to the arXiv in order to prepare for a future in which publication revenue is drastically reduced. (For example, the process of uploading a camera-ready version still involves interactions with Sheridan printing.)
Given the above, I understand why some people are considering leaving the ACM. I think that’s a mistake. I believe ACM does a lot of good, and we all benefit from a professional association unifying computer scientists. Moreover, I think leaving the ACM is not necessary. After all, the ACM is ultimately a democratic association, and open access is almost universally supported among its “rank-and file” members. So, rather than leaving, we should make a concerted effort by computer scientists across all areas to come forward with a slate of “open access” candidates for the various ACM positions. The goal would be to make sure that every democratically elected officer of the ACM supports a path toward making the digital library open access, and restructuring the publication costs in order to prepare for this future. As Moshe Vardi says, “the ACM is us“- lets make sure it behaves this way.
Update: Henry Cohn has some very insightful comments below, including why you should not use “authorizer” links.
Windows On Theory 
A couple of notes: I didn’t mention ACM’s “authorizer” links, since I’m not sure if it’s a step forward or backward from the situation where authors and conference website contain links to the arXiv version.
Also, the above discussion refers to electronic conference papers (which I believe are the majority of CS papers). Print and journal papers might have more costs associated with them. (For example, journals typically have paid administrative staff involved in managing the refereeing process, sending emails etc.., while in conferences this is done by the chair and PC.)
I consider the Author-izer links to be extraordinarily damaging for the community. They sound like a good idea at first, but in fact they substantially decrease access to the papers and increase the effectiveness of the ACM’s paywall.
The problem is that they only work if someone is accessing the paper by directly following a link from your home page, based on the http referrer field. The link no longer gives free access if it is put in e-mail or found through a web search or Google scholar (instead, it delivers people to the paywall, with no indication that a free copy could have been obtained through the author’s web page). This is a big problem, since people are overwhelmingly more likely to arrive via a search than directly. So the Author-izer service is highly misleading, since it looks at first glance like it’s about as good as being able to post the final version on your home page, but it’s much worse. If you replace the author’s manuscript with an author-ized link, then it substantially decreases the accessibility of the article, so I view this service as an extremely destructive idea. Everything I’ve said is implicit in the ACM’s description of the service (and I’ve tested it to make sure I understand how it works in practice), but they do not make the consequences clear. I consider their promotion of this service to be an attempt to trick authors into removing search-accessible versions of their papers from the web.
+1!
I recently asked the ACM Digital Library about saving more than one article at a time to a binder (for later download).
It took several exchanges but the bottom line was: Members have asked before, its up to the web portal design team and no, you can’t find out who is on the web portal design team.
My request for papers citing or cited by a paper as a download was simply ignored.
I don’t object to the ACM charging for value-added content that it generates to benefit members/subscribers. But posting papers written and edited at no expense to the ACM isn’t a value-add proposition.
It won’t simply be a matter of electing member friendly ACM officers. Those officers will need to break the grip staff has on the web and publishing processes.
Count me as interested and willing to help.
Patrick
Replacing the ACM officers with OA-friendly candidates is a great idea. Right now, the ACM is one of the worst offenders among professional societies (the American Chemistry Society is worse, but they aren’t many others). For example, the ACM has outrageous copyright policies (http://www.acm.org/publications/policies/copyright_policy), which actually manage to be worse than Elsevier’s pre-boycott copyright policies, not just the somewhat improved policies. For example, they do not allow posting to the arXiv or other repositories without special permission, and they announce that “ACM does not generally grant permission to post ACM copyrighted works on other servers acting as public repositories that compete with the ACM Digital Library.” As far as I know they don’t crack down on arXiv posting in practice, and I don’t think they could get away with it if they tried, but having policies that seemingly do not allow it is offensive, and it’s absurd to ask authors to sign forms that do not allow standard professional activities.
I really don’t understand how the ACM ended up with such a backwards position on copyright. Other professional societies (such as the American Mathematical Society) have demonstrated that reasonable policies are financially viable. Why isn’t the ACM living up to its obligations as a professional society?
By the way, I wouldn’t describe Elsevier as making math papers open access after four years, but just available for free (since they could revoke access at any time, and I’m sure they would if they thought that would put them in a position to make more money). This is the same thing the AMS does after five years. It’s not as good as open access guaranted by an actual license, but it’s certainly much better than keeping everything behind a paywall forever.
Thank you Henry for these comments, especially for noting the problems with authorizer links and correcting my mistake of saying that Elsevier makes math papers open access. I will update the post to correct this.
I do not support the extreme open access positions but I certainly agree that it would be nice if the ACM will move to a more reasonable policy. The analysis in the post is quite useful. The cost of publishing a paper which is carried out by a publisher is rather small compared to the overall cost of the paper. If we can spend more money in order to raise the quality of the final product this can overall be good. And also if we can spend less time and effort of research-mathematicians on matters that can be done by others this can also be overall beneficial. Both the open-access matter, and the precise models for publishing, look secondary issues for me, except that some proposed solutions may have (bad) implications on academics, research and researchers, which go beyond the strict publishing issue.
On another matter, I enjoyed a lot your unique game lecture on Wednesday, Boaz. (It would be nice to discuss again this conjecture on the blog…)
Thank you Gil! We already had a discussion of the UGC in this blog: https://windowsontheory.org/2012/07/31/truth-vs-proof-the-unique-games-conjecture-and-feiges-hypothesis/
perhaps the next discussion will be in your blog.. 🙂
Thank you for this post, Boaz. I have sat on the ACM Council for the past couple of years, and have not found much company when I try to express the perspective described in your post (which I share). It would be great to get more people who are concerned about these issues to run for leadership positions in ACM (or volunteer to serve on the Publications Board).
My impression is that there are several reasons for ACM’s slowness to act on this issue:
– ACM historically viewed itself as one of the “good guys” in access to publishing – with fairly low subscription rates (particularly given SIG membership, which is quite cheap). There is a perception that they are unfairly getting caught in the crossfire in the fight between the research community and commercial publishers like Elsevier, and a reluctance to acknowledge that there are good reasons for the debate moving beyond subscription rates.
– There is a misframing of the debate as being about author’s rights vs. publisher’s rights, and a feeling that it is mostly a fringe portion of the research community that cares about these rights (given that papers are available at low subscription rates). However, I think that the correct framing is society’s rights vs. publisher’s rights. As described in your post, society (eg through federal grants) pays a vastly larger fraction of the costs in producing research than publishers, and for this investment, has the right to demand perpetual access to the results of that research. This is not to say that publishing is free (I think it may be overly optimistic to expect costs to approach that of arXiv), and society needs to figure out the best way to pay for the costs, but there’s something wrong in depending on the continued goodwill of publishers to access the results of research for which the publishers have only invested a tiny fraction of the costs.
– ACM’s publications revenue (especially from the Digital Library) funds many of ACM’s other “good activities” (including substantially subsidizing the SIGs), and there is an impression that libraries will cancel their subscriptions to the DL if all papers are publicly available (even in uncollected forms, such as on arXiv). I’m not convinced that the latter is an accurate prediction, and even if it is, the risks to ACM not planning for an open-access future are even greater (as discussed in your post). Of course, ACM should to be careful in not taking risks that could cause the sudden financial collapse of the organization, but I would like to see a clear plan laid out for how ACM might transition to an open-access (or OA-friendly) future so that it can be discussed by the membership and the SIGs.
Salil
We’ve just launched a petition site to do exactly this — push the ACM towards open access! I want the ACM to be something that I can love. Help occupy it!
http://teardownthispaywall.appspot.com/
There is a petition: http://teardownthispaywall.appspot.com/
Any comments?
While I certainly support making the ACM open access, I think boycotting it is too extreme. I also don’t think ACM necessarily needs to become open access immediately, but should have a plan in place for a transition toward this goal.