Let me make one thing clear from the beginning: it was the ACM’s choice to remove my publication from their workshop proceedings. I did nothing to stop them. In fact, by waiving my copyright, I made it extraordinarily easy for them to include my work in their proceedings if they wanted.
This is my story. This was the first time I have sumbitted a paper to an ACM workshop. I have sumbitted several papers to other conferences and journals in the past without issue. For this workshop I prepared a paper entitled "Lens is to Functor as Applicative is to Biplate: Introducing Multiplate" to the Workshop for Generic Programming 2011. I went through the usual refereeing process. My paper was accepted by the program committee. I made some revisions based on the referee reports. As typical, I was asked to transfer my copyright to the publisher.
In today’s world, transferring copyright is problematic for researchers like me. We want our papers as widely read as possible in order for them to be as influential as possible. Historically, the best way to do this was to have the paper published, because this would mean that copies of our work would end up getting disseminated to university libraries around the world. Publishing is not free, but in lieu of payment for publishing, we would transfer our copyright to the publisher. However, in today’s world the best way to have my paper widely read is to submit it to an online repository, such as the arXiv, where anyone with internet access can get instant access to my work.
So as per my copyright policy, I uploaded my final version of my paper to the arXiv under a public domain dedication. When I was at Radboud University I would also submit a version to the university digital library as well. One version would be under a non-restrictive creative commons license just to cover all bases. However, the new creative commons public domain dedication seems better today and I do not have access to the Radboud University digital library anymore. Therefore, submitting to the arXiv seems sufficient for now. Because it is in the public domain, I, or anyone else, can re-archive my paper elsewhere at a later time.
Before such submission I check with the conference/journal editors to see if this is acceptable. Most journals and conferences will not accept previously published submissions. However this policy usually refers to previously peer-reviewed publications, and since the arXiv is not peer reviewed, no editor has yet had a problem with this. Most editors will tell me that publication requires a copyright transfer and the publisher might not like my final version published on the arXiv, but they maintain that this is an issue between the publisher and me. I admit that for this workshop publication I let this slip a bit. Due to the time constraints, I ended up contacting the workshop chairs after submitting the the arXiv rather than before, but, as usual, the chairs had no problem with a final copy of my paper appearing on the arXiv.
In previous publications, I print out a copyright transfer form from the publisher. Recall, that transfer of copyright is in lieu of payment for publication. It would be misleading, and possibly illegal, for me to transfer copyright to them after publishing my papers under a public domain dedication or creative commons license without notifying the publisher. Therefore, I always amend the copyright transfer agreement to make a note that I have already published my work under a public domain dedication and creative commons license and I am only transferring copyright to the extend possible (which I believe amounts to nothing). After mailing or faxing the amended copyright transfer agreement to the publisher, no publisher has yet refused to publish my work. They publish it after copy editing it, and stamp their own copyright on it. I find their copyright claim dubious; but I have no incentive to pursue the issue.
With the ACM things are a little different. Firstly, they want me to put their copyright notice onto my work myself. Secondly, they have a dubious on-line electronic copyright transfer assignment instead of a paper form. It is one thing for the publisher to put their own copyright onto my public domain dedicated work, it is another thing for me to do it myself. So instead of putting the standard ACM copyright notice into my paper, I replaced it with the creative commons public domain dedication and submitted that to the printer. Unfortunately the on-line electronic form proved to be more problematic. It is not possible to make amendments to an electronic form. (Believe me, I tried. I looked to see if I could modify the HTML, but the dubious electronic form only submits the state of checkboxes, not the text of what is agreed to.) Not really having any recourse, all I did was wait.
After I got my second reminder to fill out the copyright transfer form, I replied saying that since my paper was in the public domain, it was not necessary, nor possible, for me to transfer copyright. Then the printer asked for a copy of the permission letter from the ACM to change the copyright notice. This is when things started to take a turn for the worse. I had no such letter. So, I asked the chairs of the workshop. They had no problem with such a public domain dedication, however they do not have authorization to let me amend the ACM notice, so my request was forwarded to ACM legal where my request was promptly denied.
At this point in the story, I would like to pause and note that the ACM does have a procedure for accepting public domain papers. If I were a US government employee, I would be required by law to place my work into the public domain. The ACM has arrangements to accept publications from US government employees and uses a modified copyright notice for them. But if you are not such a government employee, well … we shall see.
I replied to ACM legal explaining that I have put a copy of my publication on the arXiv under a public domain dedication and therefore I was unable to transfer copyright to them. ACM legal replied,
The ACM still allowed me to present my work at the workshop, but they refused to allow my paper to be published in the proceedings.
Instead the chairs were allowed to insert a small note stating,
We note that one of the papers presented in the workshop is not included in the proceedings. This paper, ‘Functor is to Lens as Applicative is to Biplate: Introducing Multiplate’ by Russell O’Connor, is accessible as arXiv:1103.2841v2 [cs.PL].
As I said above, it is the researcher’s interest to have their publications as widely distributed as possible. When researchers transfer their copyright to the ACM they get a few very limited rights in return. They can post a copy of their work on their own “Home Page” and on their institution’s server. You do not have permission post your paper anywhere else, so no matter how technology changes you are stuck with using your “Home Page” (I imagine if the ACM copyright policy was formed 20 years ago you would be limited to posting on your personal Gopher page). You do not have permission to email copies of your paper to colleagues. You do not have permission to print out your paper and hand it to colleagues. Your colleagues who download copies of your paper from your “Home Page” do not have permission to print out your paper. You are allowed to make derived works of your paper; however you must include an ACM citation. You do not have permission to give anyone permission to include your paper in another publication, say of collected works in type theory, or top 10 most influential computer science papers of this century.
Why do researchers put up with transferring their copyright to the ACM? The most common answer to this is that you have to do so in order to get your paper accepted into high quality ACM conferences. However, the ACM does not make conferences high quality (at least I am not aware of any significant resources they provide, granted I have never organized an ACM conference). It is the organisers, committee, and especially the participants that make a conference high quality. I think that if we had not evolved into this situation where top conferences require copyright transfer that we would never accept it.
What shall we do about it? I was originally planning to never participate in an ACM conference again; however one participant that I talked to suggested that I keep doing what I am doing. I can continue to post my final version of my publications on the arXiv and not transfer my to the ACM. The ACM can continue to refuse to put my publications in their proceedings. However, I as long as I can continue to present my work at their workshops and conferences, although not ideal, it is good enough for me. I will continue to list my papers on my CV as peer reviewed. People will still be able to find my publications from my DBLP page. And my publications will be part of the long term CoRR archive, an archive that is likely more resilient that even the ACM’s digital library. If you care to join me, then pretty soon the ACM will have no more papers in their proceedings. Maybe that will make them rethink their policy.