March 30, 2012
Tonight, I sent my submission to Research Councils UK in response to their call for comments on the recently issued docment RCUK Proposed Policy on Access to Research Outputs. I am now posting my comments publicly. I urge you all once more, please send your own comments to firstname.lastname@example.org with the subject “Open Access Feedback”. They do not need to be as long and detailed as mine: I am sure they would welcome short-and-sweet comments!
I enthusiastically welcome the proposed changes in the the RCUK Policy on Access to Research Outputs, recognising that this policy’s more timely availability of access to research and the more liberal licensing terms will be good for research, for industry and for society in general. Some minor harm to the businesses of subscription-based publishers is a regrettable possible side-effect, but that harm — if it is real, which has not been demonstrated — is greatly outweighed by the benefits that will arise from the adoption of the new access policy.
I will comment on each of the changes individually.
(2) What do the Research Councils mean by Open Access?
“The existing policy will be clarified by specifically stating that Open Access includes unrestricted use of manual and automated text and data mining tools. Also, that it allows unrestricted re-use of content with proper attribution – as defined by the Creative Commons CC-BY licence.”
It is very heartening to see this clarification, especially as certain publishers seem to be using the phrase “open access” extremely loosely to refer to any articles to which any kind of access is provided. The definition used in the RCUK policy is compatible with that of the Budapest Initiative which first defined the term in 2002.
Particularly welcome is the clarification that “open access” may not exclude commercial use — a clause that is sometimes adopted by authors or journals that hope to gain financially by forcing commercial organisations into a royalty agreement, but which almost invariably simply prevents the research from being used.
(3) How is a Scholarly Research Paper made Open Access?
One of the two methods by which a paper can be made Open Access is given as follows:
“The version of the published paper as accepted for publication … is archived and made accessible in an online repository … access may be restricted to comply with an embargo period imposed by the publisher.”
This wording could be improved by clarifying that the deposit itself is to be made as soon as the paper is accepted, but with an embargo period before the deposited manuscript is made open access. Most modern repositories support such “dark deposits”, with the open-access date specified at deposit time so that no further human intervention is required six months later when the embargo expires.
This approach would have several advantages: first, deposits would be made when the project is fresh in the author’s mind. Second, article metadata (though not full text) would be available from acceptance time, speeding recognition of the paper and application of the research. Third, potential readers who discover metadata in advance of embargo expiry will be able to obtain copies direct from authors.
(4) What do journals need to do to be compliant with Research Council policy on Open Access?
“The existing RCUK policy on access to research outputs does not state specific criteria to be satisfied for a journal to be recognised by the Research Councils as ‘Open Access Policy Compliant’. The revised policy therefore introduces such criteria.”
This is an important and very necessary change, in light of the variety of ways the term “open access” has been abused.
Although the Directory of Open Access Journals (http://www.doaj.org/) lists over 7000 “open access” journals, specific licensing terms are specified only for a very small proportion of these, so that users cannot easily tell what rights they have regarding articles from most listed journals. A more explicit list of True Open Access journals will be helpful, especially to text/data mining projects. I hope that RCUK will either establish such a list, or (better still) work with DOAJ to add an RCUK-compliance field to its database.
(5) What Research Outputs will be covered by Research Council Policy on Access to Research Outputs and where should they be published?
No comments other than agreement.
(6) When should a paper become Open Access?
“In future, Research Councils will no longer be willing to support publisher embargoes of longer than six […] months from the date of publication, depending on the Research Council.” [for councils other than AHRC and ESRC]
This is definitely an important step in the right direction.
However, I question whether any embargo period at all is acceptable for research funded by the public. I understand that the six-month period is a compromise in hope of appeasing publishers, but the core point here is that the Research Councils are not beholden to publishers but to the British public. Their goal is to obtain the best value in return for taxpayer investment in research, not to perpetuate the business model of old-school publishers.
I would therefore support a no-embargo rule, whereby final manuscripts could be posted to repositories as soon as they are accepted (i.e. even before publication). Publishers that are unhappy with this arrangement would be free not to accept manuscripts submitted under these terms, and to seek submissions from elsewhere.
(7) How is Open Access paid for?
“Research Council grant funding may be used to support payment of Article Processing Charges to publishers.”
This principle is a good one. However, there are practical difficulties in estimating at the beginning of a project how much money to request for publication fees when it is not known how many papers will proceed from a project, what journals they will be submitted to, or whether they will be published during or after the lifetime of the project.
For this reason, rather than including publication fees in grants, I would favour the establishment of a separate pot of funds dedicated to supporting publication of RCUK-funded research whether during or after any given project.
(8) Acknowledgement of funding sources and access to the underlying research materials
“Research papers [must include] a statement on how the underlying research materials — such as data, samples or models — can be accessed.”
Requiring this to be explicitly stated is a valuable addition which is cheap to comply with.
But I am disappointed to find so large a loophole as “The underlying research materials do not necessarily have to be made Open Access”. While understanding the need for some datasets to remain private (e.g. patient records and other personal medical data), I would prefer to see such exceptions listed, with a clear expectation that datasets not in one of the exception categories should be made open access. The motivation for this change is the same as that for the whole policy: that free availability of data, like research, accelerates both further research and commercial applications, to the benefit of the public.
(9) Implementation and compliance
No comments other than agreement.
March 30, 2012
Research Councils UK is the aggregate of the UK’s seven research councils, which makes it overall the most important and influential funding body for science in Britain. A few days ago, they released a draft of their new open access policy, and they are soliciting comments now. Comments can be from anyone — individuals or groups, British or overseas — like the recent OSTP Request For Information in the States which we have to assume was influential in the defeat of the RWA.
What does the new draft policy say? I urge you to read it yourself: its only six pages, and they are very clear. To quote from the document’s introduction:
Key differences with the current policy include:
- Specifically stating that Open Access includes unrestricted use of manual and automated text and data mining tools; and unrestricted reuse of content with proper attribution.
- Requiring publication in journals that meet Research Council ‘standards’ for Open Access.
- No support for publisher embargoes of longer than six months from the date of publication.
Unsurprisingly I am very much in favour of the proposed changes to the RCUK policy, and I will be a making a lengthy submission commenting on individual changes.
But any comment is significant — even you’re just writing to say “I approve of the policy changes”, or “I recommend a 12-month embargo period instead of six” or indeed “open access is a silly fad, this policy in unnecessary”. The point is that RCUK want to hear your opinion. Your voice matters.
So please email your comments to email@example.com with the subject “Open Access Feedback”. I am told (by Cameron Neylon, who heard it from a colleague who’d had it from a friend of his) that the deadline is Thursday 5 April, but since that is hardly a definitive source, I recommend getting your submissions in as soon as possible. (You can be sure that the publishers will be doing so.)
March 29, 2012
Just a quick note that I’m the interview subject in the P.S.I.O.N podcast this week. P.S.I.O.N is the Paediatric Surgery International Online Network — an area far outside my expertise, but of course what we talked about was open access rather than paediatric surgery.
You might be forgiven for thinking that I’ve said all I have to say on open access by now — to be honest, I sometimes think so myself — yet when I listened back through the podcast I found that I surprised myself with some of the thoughts that came up. In fact, call me narcissistic, but I thought it was really interesting :-)
That’s thanks largely to excellent questioning from the interviewer, Eve Macharia. Eve is a research associate at UCL, a paediatric surgeon in training, and managing editor of P.S.I.O.N, which is both a network and an open access journal running on WordPress.
March 29, 2012
By one of those happy coincidences that you sometimes get, today saw the publication of not one but two dinosaur ontogeny papers: this morning I was sent a copy of Woodruff and Fowler (2012) on ontogenetic changes in the bifid spines of diplodocoids, and tonight I was alerted to Werning (2012) on Tenontosaurus growth trajectories based on osteohistology.
It’s interesting to compare them. The obvious conclusion is that, while sauropod vertebrae are intrinsically better than ornithopod long-bones, the latter make much, much better subjects for ontogeny studies.
The problem that Woodruff and Fowler have is that they’re working from a small selection of mostly isolated elements, nearly all of them damaged by breakage and/or crushing, and uncontrolled for serial position beyond some basic binning. As a result, the taxonomic identifications are really rather arbitrary and unsupported — as they say (p. 2), “for the purposes of this study, original taxonomic designations were not reexamined” — and indeed re-examination would not necessarily help much.
As a result, they are left with rather a circular argument, as follows:
- Small specimen X is Suuwassea.
- Small specimen Y is assigned to Apatosaurus, because someone once said so.
- But Y has a less bifid neural spine than adult Apatosaurus.
- So spine bifurcation increases through ontogeny in Apatosaurus.
- So small specimen X belongs to Apatosaurus, too.
- => Suuwassea is Apatosaurus [Note: I stupidly wrote “Diplodocus” here in the first posted version. Now corrected to Apatosaurus.]
I don’t find this at all convincing. (Neither does Matt: we discussed this briefly today, and at more length at the Bonn workshop where Cary presented this work.) Leaving aside the observation that the conclusion fits in neatly with the Horner Lab’s ongoing everything-is-just-a-Triceratops-growth-stage project, there isn’t really much that Cary could have done differently here: the necessary specimens (i.e. multiple near-complete associated individuals of unambiguous taxonomic identity) just don’t exist.
(Mind you, figure 7A is about the least convincing evidence of bifurcation I’ve ever seen.)
Ornithopod long bones
By contrast, Sarah Werning has a much better story to tell, largely just because she’s working with much better material. Tenontosaurus is known from many complete and near-complete individuals, there is no real uncertainty about the material all belonging to the same taxon, and the elements (long bones) are much less subject to crushing and breakage than vertebrae.
Of course it also helps that Sarah has done astonishingly careful, detailed work. I don’t think I’m giving much away when I say that this paper has been many years in the making — as the Acknowledgements say “This work was completed in partial fulfillment of the requirements for the Master of Science degree, Department of Zoology, University of Oklahoma.” And it’s been a looong time since Sarah was at Oklahoma. But I get the sense that this reflects taking the time and putting in the effort to Get It Right, rather than standard-issue procrastination.
Perhaps the most impressive part of this paper, though, is methodological:
High-resolution histological images of the cross-sections are digitally reposited online for scholarly use at MorphoBank (http://MorphoBank.org), project p494; see Table 2 for a list of slides and accession numbers. Digital images larger than 25,000 pixels in either dimension were digitally scaled (reduced to 17,000–20,000 pixels in the larger dimension) to allow processing on MorphoBank and because most image editing software does not support editing of gigapixel jpg files. These edits were made after scale bars had been added. Images in full resolution can be obtained from the author.
As you can see, these are ridiculously high-resolution images. By making them freely available, Sarah is doing everything she can to enable others to replicate or correct her work — or to use the data for other purposes that she’s not yet thought of. This is a big win for the progress of science, and I want to publicly congratulate Sarah for doing it. I would love to see this become standard behaviour.
(It’s a bit strange that the paper links onto the the MorphoBank home page rather than directly to the Tenontosaurus project. Here you go.)
PLoS ONE vs. Journal of Morphology
A final thought, before I finally go to bed. When I saw Cary’s paper in the Journal of Morphology I thought to myself, “good for him, he’s got that into a really good journal” — which is true. But then when Sarah’s came in, I found myself comparing PLoS ONE. Here’s what I came up with:
Access. A no-brainer: PLoS ONE wins hands down, being open-access while J. Morph. is paywalled.
Charges. The flip side of the first category: J. Morph. wins because (as far as I can tell from the Author Guidelines) there is no publication charge. PLoS ONE charges $1350.
Length. Another PLoS ONE win, because it imposes no limits whatsoever on length, figure count, etc. Cary’s paper is no lightweight at 11 pages, but PLoS’s liberality with page-space means that Sarah’s is well over twice as long at 25 pages. Now of course not all papers need to be long; but for those that do, cutting to a journal’s length limit is a painful and stupid process.
Image Colour. Again, PLoS ONE wins — it’s very rare these days to see a greyscale specimen image in a PLoS journal. But to be fair to J. Morph., the author guidelines do say: “All color figures will be reproduced in full color in the online edition of the journal at no cost to authors. For the printed version free color figures are at the editors discretion.” So it looks like Cary and Denver dropped the ball on this.
Image Resolution. No question. PLoS ONE wins by a mile. Click through the two representative images above: the PLoS one has 3.5 times as many pixels, and that’s after I added a 10% wide margin around the J. Morph. image. And if you want to use these illustrations as basis for new images, remember PLoS also lets you download the original full-resolution submitted image in a lossless format.
DOI Resolution. When I added the references below, I noticed that the DOI for the J. Morph. article doesn’t resolve yet, which is careless. PLoS ONE wins in this category, resolving just fine.
Impact Factor. Yes, it’s a stupid number and we all hate it. But some people still seem to take it seriously, so we may as well look. And PLoS ONE wins with 4.411 to J. Morph.’s 2.087 — more than double.
Putting it all together, based on the seven categories that I evaluated here (and no doubt I missed some), it looks like the only reason to go with J. Morph. ahead of PLoS ONE is to avoid the publication fee. By this point, PLoS ONE has made itself an absolutely mainstream journal for palaeo, and the obvious first choice for most projects.
- Werning, Sarah. 2012. The ontogenetic osteohistology of Tenontosaurus tilletti. PLoS ONE 7(3):e33539. doi:10.1371/journal.pone.0033539
- Woodruff, D. Cary, and Denver W. Fowler. 2012. Ontogenetic influence on neural spine bifurcation in Diplodocoidea (Dinosauria: Sauropoda): a critical phylogenetic character. Journal of Morphology, online ahead of print. doi:10.1002/jmor.20021 [Direct Link, since the DOI doesn’t seem to work.]
A few weeks ago, I noted that the new journal Biology Open, which had just published its very first issue, had made the unfortunate choice to use the Creative Commons Attribution-NonCommercial-ShareAlike (CC BY-NC-SA) license. Although this licence does offer a lot of freedom, it’s too restrictive to be properly described as “open access”, a term originated in 2001 by the Budapest Open Access Initiative which defined it as follows:
By “open access” to this literature, we mean its free availability on the public internet, permitting any users to read, download, copy, distribute, print, search, or link to the full texts of these articles, crawl them for indexing, pass them as data to software, or use them for any other lawful purpose, without financial, legal, or technical barriers other than those inseparable from gaining access to the internet itself. The only constraint on reproduction and distribution, and the only role for copyright in this domain, should be to give authors control over the integrity of their work and the right to be properly acknowledged and cited.
I wrote to the editors, as I noted last time, and received a polite reply from Rachel Hackett, the managing editor. After a little more back and forth, she sent me the following statement from the Company of Biologists with permission to publish it on this blog:
“The Directors of the Company of Biologists, who are active scientists, reviewed the licence under which Biology Open publishes its articles both before and after launch.
We are aware of the implications of the different licence types and it was decided that the CC BY-NC-SA licence was appropriate to ensure that the rights of both author and publisher are protected. Those wanting to use our content for commercial purposes can still ask us for permission.
The Company of Biologists is a not-for-profit Charity that finances journals, supports meetings and awards travelling fellowships.”
So there are a few things to say here:
1. It’s their journal, and of course they have every right to publish it under whatever terms they want.
2. It’s a pretty open licence, allowing the articles to be freely read by researchers and to become input to further (non-commercial) research. So credit is due, and I hope no-one at Biology Open feels I’m picking on them. But:
3. It’s a missed opportunity. Heather Piwowar said this rather well in a recent comment: “We do basic research not only to know more, but to do more”. Non-commercial licences impede the use of research, and that’s not to the benefit of wider society. (I won’t labour this point now, because I’ll have more to say on non-commercial clauses in a subsequent post.)
The upshot here is that, because no-one really knows what “commercial” means, there will be a chilling effect on using Biology Open papers for anything other than human reading. For example, consider a company wanting to run text-mining software across a corpus of papers to discover chemical reactions. They will know they can do this with, say, PLoS articles (which are CC BY), but will have to skip Biology Open if they want to stay on the safe side. No-one gains by this exclusion — certainly not the journal or its publisher.
March 26, 2012
My new piece is now up the LSE Impact Blog — in which I recognise that it’s a mistake to think of Elsevier and other for-profit barrier-based publishers as evil. The money quote:
Talk of such publishers being “evil” is really misplaced. They do what they do. It would be more accurate to call them “blind” or “unthinking”. When they fight tooth and nail to prevent open access, they are no more being evil than a shark is when it attacks its prey; no more evil than a brick wall across a motorway.
By the way, if you read it, do leave a comment; and if you like it, pass the link to your colleagues. It’s great to get discussion going on these pieces, bringing the issues to a new audience each time. (I realise that for SV-POW! readers, the core issues are now well rehearsed; but we need to remember that even now only a very small proportion of academics recognise the importance of open access.)
A short one, because I’ve been commenting on other people’s blogs a lot recently (Scholarly Kitchen, Open and Shut, The Scientist) and it infuriates me how hard it is get a good back-and-forth discussion going in those venues.
The contrast of course is with SV-POW! itself, where we’ve often had excellent, busy, informative comment-threads (example 1, example 2, example 3) that have resulted in us learning a lot from our commenters. So why is it that some blogs’ comment streams are lively and productive, whereas others are relatively sterile?
Here’s how to do it right.
1. Turn on comments. It should be no-brainer, I know, but it always brings me up short when I find a blog post with commenting disabled. When a blog keeps doing it, it’s one of the most likely factors that will make me give up on the blog altogether. If it’s just one person preaching, it’s much less interesting than a discussion.
2. Do not require registration. This one, too, should be too obvious to need stating. Notoriously, the Scientific American blogs require you to register before you can comment, hence the fact that the SciAm incarnation of Tetrapod Zoology gets less than half as many comments as the old Science Blogs version used to get. And that in turn is why even now, eight months after Darren stopped posting at the old site, it’s still Google’s top hit for “tetrapod zoology”.
Science Blogs lets you comment just by filling in your details on the comment form, and of course browsers do that for you, so the process is painless. “But registration is fast and free”, say the SciAm people. Doesn’t matter: it’s a road-bump. At any given moment, the point someone wants to make in a comment is probably not enough to push them over that bump. So don’t make them do it. They just won’t.
There are various open-ID schemes that different blog platforms support. Most people by now have at least one of a Google, FaceBook or Twitter account. If you let people use those as their commenting identity, the problem goes away. If your blogging platform doesn’t support this, you might have chosen the wrong platform. (Why SciAm decided they needed to build their own, I can’t begin to imagine.)
3. Do not use CAPCHA. I can’t tell you how much I loathe this. I type a comment, try to submit it, and am met with this brain-damage:
What the heck does that say? Is that “ri” or “n”? Is that a “t” or and “f” just before that other thing that might be “1” or “l”? And why are you putting me through this when I am trying to contribute to your blog?
Again: people will just shrug and walk away.
Bloggers who use blogger.com tell me that they have no choice about this, that the platform imposes it. If that’s really true, then ditch blogger.com. It’s just broken.
4. Do not moderate comments. This is the potentially controversial one, because up till this point I don’t think I’ve said anything that anyone reasonable would disagree with. But this one is counter-intuitive. When you start a blog, the natural thing is to want to feel that you’re in control of it, and that means controlling what can be posted there. But that’s a mistake. Moderation means that people can’t see their own comments, which is alienating; but more importantly, it means other people can’t see them, which in turn means that all discussion grinds to a halt until such time as you happen to moderate.
What that means is that the site is only really alive when you’re at the keyboard, constantly checking your inbox, so that you notice moderation requests as soon as they come in. It means you’ll never have the experience of waking up in the morning and finding that a discussion has broken out on your blog.
But what about spam? On a good platform, it’s not a problem. Since we started SV-POW!, 6,539 comments have been posted, and 3,552 spam comments have been automatically detected and help for moderation. My and Matt’s manual moderation of those suspected-spam comments shows that detection has been 99.92% accurate: there have been only three false negatives in five and a half years. There have been 63 false positives, i.e. comments that looked like spam but weren’t. Those were held for moderation, and passed.
So. You don’t need to moderate to filter spam, and you don’t want to moderate to control discussion. Just open it up.
5. Allow subscription to followups. Some platforms let commenters tick a box or click a link to ask that they be emailed when someone posts a followup comment. That’s a very valuable feature, because it makes people aware of how the conversation they’ve joined is progressing, and gives them a chance to respond.
6. Reply. A very obvious one: engage with the comments yourself. You want to be a part of the conversation, as well as having the privileged position of setting the agenda.
7. Oh, and write something interesting. Your posts catalyse the comments. Once a comment stream is up and running, it has its own momentum, but it will only get started if you give it a push. Now I am not suggesting that you deliberately set out to write controversial articles, or really that you set out to write anything other than what interests you. There is really no point in writing about anything that doesn’t interest you — if you’re not enjoying your blog, then no-one else will, either.
Now that I come to review, I almost wish I’d said less — because one of these seven points is the burden of my message, and that one may have got a bit lost in the middle. It will come as no surprise to anyone that if you want fruitful discussions in your blog’s comments, you need to turn on comments, avoid speed-bumps like registration and CAPCHA, let people subscribe to followup comments, reply, and write interesting articles. The only remotely controversial part is do not moderate comments, and that’s the main point I want to make.
So here it is again: do not moderate comments.