- Sending quantum information down channels which cannot convey quantum information: A perspective in Science. A copy is available here.
- Quantum computing as free falling: A Science perspective on quantum computation as geometry. A copy is available here.
- Quantum information can be negative: an accessible description of my paper with Horodecki and Winter in Nature
- Copyright transfer and physics journals: the American Physical Societ has accepted our proposal to allow authors to freely make available parts of their work.
- Administration Free Thursdays: every Thursday, no grant proposals, no fellowship applications, no reports, no booking of speakers or claiming of expenses. We just do physics.
Traditional journals and copyright transferMay 11th, 2007
Updated (October 1st, 2008): The American Physical Society has accepted one of our two proposals!
Updated (March 10th, 2008): Physicists who submit their papers to online archives now have a tough choice ahead of them following recent licensing changes. Authors submitting to the arXiv now have the option to release their work under one of four licenses which would allow them or the public to contribute parts of the work to such efforts as Quantiki and SklogWiki (which are a little bit like Wikipedia, but more current and specialised). This is great news, as it gives authors and the public more freedom to innovate. Indeed many authors have chosen to release their work to the arXiv under the more permissive licenses, and have then submitted their paper to a traditional journal
However, this may clash with the current policy of the American Physical Society (APS). Many of us hope that the APS's policy will be clarified to keep in step with changes to the arXiv, and innovations in online publishing. The APS currently allows authors to submit their papers to the arXiv, but they do not specifiy which license is to be used. They have accepted some papers which have been released under the more permisive licenses, but have revoked publication in other cases. They have generally been good about giving their authors more freedom, so I hope they will allow authors to submit their papers to the arXiv under a license of the author's choice.
Traditional journals and copyright transfer
"In the firm belief that an understanding of the nature of the physical universe will be of benefit to all humanity, the Society shall have as its objective the advancement and diffusion of the knowledge of physics."
-- mission statement of the American Physical SocietyTechnological advances have led to a vast array of tools that scientists can use to communicate their ideas. These tools include open access journals, online archives, paper rating websites, science blogs, quantum blogs, open courses, free universities, and open encyclopedias such as Wikipedia. Of particular interest are more specialised encyclopedias such Quantiki (a Quantum Mechanics and Information Wiki initially set up at the initiative of the University of Cambridge), Qwiki at Caltech and the statistical mechanics wiki SklogWiki. Not only does this enable scientists to better share ideas with each other, it creates a commons of scientific information that is freely available to the public. Some of these projects are advanced, while some could obviously use some help. The public funds our research, and they should benefit from its fruits. Likewise, scientific publishing will no doubt change dramatically, and we should ensure that we will be able to use our research in any new forums which will arise. We would be crazy to cut ourselves off from any new advances without sufficient reason. The most well known example of a commons of free content is Wikipedia -- a recent survey of scientists by the journal Nature found that seventeen percent of Nature authors consult Wikipedia weekly. Nature has urged scientists to increase their contributions to Wikipedia, thus enlarging and strengthening public content.
Recently, one of the APS journals, Physical Review Letters (PRL), rescinded their previous acceptance of two papers by myself and coauthors [OW][OSW]. This occurred because we wanted the option to contribute parts of our paper to the intellectual commons. While PRL had allowed such options in the past, and initially agreed that this was permissible, their current policy inadvertently prevents authors from posting their own figures on Wikipedia (for example). This means that when you submit your paper to the arXiv, you need to be very careful which of the four licenses you choose. If you check any but the first box, you may find yourself in the same position as us.
Before PRL will publish an article, the authors must transfer copyright by signing this form. Once one does this, the author is no longer the owner of the work -- the APS can theoretically sue for copyright violation should you use your own words in a manner not covered by the transfer of copyright form. To be fair, they will almost certainly look the other way. In fact, as will become clear, most physicists routinely do things which violate the transfer of copyright agreement. This creates the unfortunate situation where those who take the transfer of copyright agreement seriously are penalized. Do we want to create a situation where we advocate scientists sign a contract they know they will have to violate?
There are, admittedly, reasons why the APS feel they need the transfer of copyright (more about this in a moment). However, in the past, one could negotiate an exception. For example, Bill Unruh had such an exception on his papers, and here is a copy of similar exception we were given by another journal, Communications of Mathematical Physics. It is virtually identical to the one we asked for from PRL which resulted in them rescinding acceptance of our papers. For-profit journals tend to have more restrictive policies than the APS, while the Royal Society has a progressive policy when it comes to author rights. You just grant them a license to publish, and retain the copyright. You can release your work under a creative commons license, but are charged a fee which is so high that in my view, it's prohibitive. As an aside, their policy on open access is not as good as the APS since you must wait a year before an author can put the APS-edited version of the paper on their website (although the author can post the pre-print version at any time). The ACM and Springer have allowed use of a license rather than transfer of copyright if requested. Note that US government employees cannot transfer copyright by law, and therefore, the APS and other journals have to anyway make an exception in such cases and effectively rely on a license rather than copyright transfer.
Why does PRL require its authors to transfer copyright? According to their outgoing editor in chief
"Although we hold copyright on articles, we do so only to ensure that we are able to do what we wish with them. In practise, authors retain all the rights that they would have were they themselves to hold copyright."So, for example, if instead of transferring copyright, authors granted them a license to print the article, then it might be unclear if PRL would also be allowed to put the article online. Or if in the future PRL wants to do something else with your article, they want the freedom to do it without having to ask your permission again.
Fair enough. But physicists also want the freedom to innovate, and I imagine that the APS is equally concerned about its author's rights and freedoms, as its own. Surely it is possible to craft a document which preserves the interests of both parties -- which gives both parties and the public the freedom to innovate? PRL believes that in practise authors should retain all the benefits of copyright holders. It would therefore be desirable to amend their copyright transfer form to ensure that their intent is realised.
PRL might be concerned about their ability to recoup their costs. However, in this case they are correct not to be concerned -- giving authors more rights is not going to have any significant impact on their finances. It is extremely unlikely that a library is going to cancel their subscription to PRL because parts of some articles are on Wikipedia, especially when the full article is already available on the arxiv at no cost. In fact, the availability of free copies of papers on the arxiv has not resulted in a loss of revenue. The further use of materials, if properly referenced, is more likely to increase demand for the original article rather than decrease it. One can easily imagine that someone reading Wikipedia may be motivated to check out PRL if they see parts of a PRL article there. There is no reason for PRL to feel threatened. This should all be viewed as a wonderful opportunity to broaden the scope of dialogue between scientists and the public. It is not something to be feared.
I believe the APS will do what they can to increase the possibilities available to scientists because they are responsive to their members. Indeed the transfer of copyright that the APS uses is an evolving document. For example, when preprint archives became available, the APS changed their copyright document to allow authors to post their papers there. But the current transfer of copyright was drafted before the rise of Wikipedia and is simply no longer suited to current realities.
Before turning to possible ways to update the APS transfer of copyright, it may be useful to clarify some copyright issues.
Currently, when one transfers copyright to PRL, one reserves certain rights. In particular, the right to post a version of your paper on a preprint archive and the right to post a version of your paper on your webpage. As noted, the APS are quite good in this regard and this has not resulted in any loss in revenue. However, the transfer of copyright does not allow you to do many things which are of benefit to science. For example:
- post a figure or parts of your paper (even if rephrased) onto an open encyclopedia such as Wikipedia or Quantiki which requires that your work be submitted under the GNU Free Document License (GFDL). In particular, I wanted the option to rework parts of a paper into a tutorial for Quantiki
- give a talk derived from your paper for which you receive an honorarium or at a conference which charges a fee, like the APS March meeting, especially if the talk is recorded.
- create a condensed or expanded version of your paper for a conference proceeding (since virtually all of them are published in commercial journals)
- use parts of your work in a book (if distributed by a commercial entity)
In the United States, "derivative work" is defined in 17 U.S.C. § 101:
A "derivative work" is a work based upon one or more pre-existing works, such as a translation, musical arrangement, dramatization, fictionalization, motion picture version, sound recording, art reproduction, abridgment, condensation, or any other form in which a work may be recast, transformed, or adapted. A work consisting of editorial revisions, annotations, elaborations, or other modifications which, as a whole, represent an original work of authorship, is a "derivative work".As you can see, this covers a lot of what physicists do, and it is forbidden under the transfer of copyright agreement. The issue for Quantiki and Wikipedia (and many projects which use the GFDL), is that in order to license anything based upon one of your papers, one needs to be the copyright holder (or have the permission of the copyright holder). In some situations, even posting a modified figure into a Wikipedia article constitutes the creation of a derived work.
You could use a figure in a Wikipedia article before you sign the transfer of copyright, but not after. Unfortunately, if you post parts of your paper to Wikipedia (or use one of the creative commons licenses when you submit your paper to the arXiv) before you transfer copyright to the APS, it is unclear whether they will still publish your paper. Currently, they will consider it prior-publication. It is unclear why this should be the case -- after all, if they are only concerned with the freedom to innovate, they will still be the copyright holder with all the freedom that this entails.
Perhaps they are afraid that once parts of your work end up as part of the creative commons, there is no telling where they will end up (including commercial use). However, the unlikely event of commercial use seems like a moot point, since the GFDL requires that anyone re-using your work is also required to give it away for free. It is hard to imagine that anyone could profit off a Wikipedia entry, when it is required to keep the free version available. Interestingly, here is the license that one originally granted to the arXiv when submitting a paper:
I grant arXiv.org a perpetual, non-exclusive license to distribute this article.It is far more permissive than the GFDL!
So what should be done?
At least two suggestions have been put forward. The first one involves granting the APS a license rather than transferring copyright. The other, written by Bill Unruh, involves the APS granting an exception for derived works. Update: The arXiv has made room for more options -- namely, that you have the option to choose which license your paper appears under in the arXiv, and the APS should not consider this pre-publication. Bill Unruh and I have submitted these proposals to the Publications Oversight Committee, and a copy of the proposals are available here
First the license: I believe that if we want to maximise the rights of both parties to innovate as much as possible, then the most natural thing to do would be for the author to give a highly permissive license to PRL to make use of their work. Thus we should not only give them the right to publish our articles in print and on the web, but also to publish our content in any form which exists or will exist in the future. This would appear to satisfy the stated intent of the transfer of copyright. It is also clear from previous experience with the arxiv that this will not result in a loss of revenue for PRL. On the contrary, the increased exposure is more likely to increase the revenue of traditional journals.
Here is some standard language that could be used in this regard:
the non-exclusive right throughout the world to edit, adapt, translate, publish, reproduce, distribute and display the Article as a whole in printed, electronic or any other medium and format whether now known or yet to be developed.The journal might also want the author to give them the authority to legally pursue violations of your copyright. They would anyway own the copyright to the APS-prepared version (i.e. their changes and layout), as well as the collection as a whole, and would be able to take legal action against anyone selling copies of their journal.
One might be concerned that if authors retain copyright, they might submit their paper to more than one journal, or rewrite and reuse their article without adding anything new. This is a reasonable concern, and indeed there are a few scientists who rewrite the same article many times or rewrite an article for a conference proceeding without adding any value to the new version.
However, the re-hashing of old papers is severely frowned upon and is extremely rare. No one wants to get a reputation as a scientist who simply reheats cold results; people will stop reading their papers. Copyright is not needed to stop gratuitous re-use of article text. The scorn of the physics community is surely punishment enough to discourage this activity.
On the other hand, modest re-use of figures, some text and equations is fairly common. About ten percent of papers in the arxiv contain self-copying. Clearly many authors ignore the transfer of copyright form! Indeed, strict adherence to copyright would impede much of physics. The APS knows this, and it is probably why they let most authors ignore the transfer of copyright form. However, failure to defend one's copyright can invalidate a later claim to copyright, so it is actually in the best interests of the APS to explicitly allow this sort of copying, rather than simply look the other way.
Bill Unruh has suggested another approach, namely the insertion of a clause which allows authors to reuse up to 50% of their material provided that 10% of the material is new. He has kindly provided me with a copy of the letter that he independently wrote to the APS on this. In addition to his proposal, the letter also contains some interesting legal background.
The third option would probably be ideal for all parties. Namely that the APS will not consider it prior publication if an author chooses any of the licenses available when they submit their paper to the arXiv. This is certainly the simplest solution, and perhaps the only way to account for the fact that many authors will want to choose the licenses which grant more freedoms. Already, Physical Review A and other journals has published a number of papers which have been released in the arXiv under some of the Creative Commons License. This proposal should be implemented in conjunction with one of the first two proposals, since some authors may not feel comfortable releasing their work under a Creative Commons license.
A few researchers in quantum information theory have signed a letter supporting the review of copyright policy currently being undertaken by the Publications Oversight Committee, and the Unruh proposal in particular. It is available here.
If you want to support the efforts, expecially if you are a member of the APS please drop me
an email. Likewise, if you have further questions about this page please do not hesitate to
contact me at: J.Oppenheim@damtp.cam.ac.uk .