There’s an awful lot of talk about “predatory open access publishers” recently. So much talk that I can’t help wondering whether the phrase is being pushed by barrier-based publishers in another attempt to smear open access. (Hey, they have previous.)
Anyway, for anyone who is worried that they might be tricked into giving their work to one of these low-quality predatory publishers that accepts anything and only cares about the fee, here is my guide to avoiding this scenario.
So. Imagine you have an article ready to go, when you receive an invitation to submit it to a journal that you’ve never heard of before. How do you decide whether to send it to them?
Do not send it to them.
There are probably many ways of getting a “90% complete” paper finished and ready for submission, but here’s the way that works for me. (It’s working for me right now: I’m in the middle of the process, and broke off to write this just for a a break.)
You will need:
- A printed copy of your manuscript
- A red pen
- A CD of Dar Williams songs that you know inside out
- A bottle of red wine
- A bar of white chocolate (optional)
Take the printed copy of the manuscript. read it through, with the Dar Williams CD on in the background. Every time you see anything you don’t like, scribble on the printed copy with the red pen. It might a typo, a misspelling, an infelicitous phrasing, a missing reference, a taxonomic name needing italics; or it might be something bigger, like two sections that need to be swapped.
Do you really need a printed copy for this? YES YOU DO! Can’t you just do it on the screen? NO YOU CAN’T! For one thing, you’ll keep breaking off to read email, which is a complete killer. For another, you’ve been working on this manuscript on screens for months already. Your poor brain is inoculated against its on-screen appearance. You need the mental jolt that a shift of format gives you. And you need the freedom to scribble. When I do this, I often write in suggestions to myself of what alternative wording to use, but I feel free to ignore them when I come to make the edits.
Do you really need a Dar Williams CD? I am prepared to concede it doesn’t necessarily have to be Dar Williams. But it does need to be something that you know so well that it won’t surprise you, it won’t grab your attention away from the work you’re doing. Much as I love Dream Theater, their music is really not the way to go for this. What you want is music that with keep feeding you without distracting you.
Do you really need the red wine and the white chocolate? Perhaps not, but you don’t want this to be a boring, unpleasant process, do you? Treat yourself. (DISCLOSURE: I have moved on to beer.)
As soon as I’m done posting this, I’ll be going to Step 2, which is to go through the manuscript, making edits on the master copy. Most of them are trivial to do. A few are going to need real work. For these, I just leave a marker in the master copy, “###” and a note saying what needs doing. I will later search for these and do the work. But not tonight.
The goal of this process is to capture all the information that you wrote on the printed copy, so that you can throw it away and move on with your life.
That’s it — it’s all you need to do. For the record, I expect to submit in the next three or four days.
March 11, 2013
For a paper that I and Matt are preparing, we needed to measure the centrum length of a bunch of turkey cervicals. That turns out to be harder than you’d think, because of the curious negative curvature of the articular surfaces.
Above is a C7 from a turkey: anterior view on the left; dorsal, left lateral and ventral views in the middle row; and posterior on the right. As you can see from the anterior, dorsal and ventral views, the anterior articular surface is convex dorsoventrally but concave transversely; and as you can see from the lateral view, the posterior face is concave dorsoventrally and convex transversely.
This means you can’t just put calipers around the vertebra. If you approach the vertebra from the top or bottom, then the upper or lower lip of the posterior articular surface will protrude past the centre of the saddle, and give you too long a length. If you approach from the side, the same will happen with the left and right lips of the anterior articular surface.
What are we trying to measure anyway?
But this raises the question of what it is we’re trying to measure. I said “we needed to measure the centrum length of a bunch of turkey cervicals”, but what exactly is centrum length? Why shouldn’t the upper and lower lips of the posterior articular surface count towards it?
What does centrum length mean?
The problem doesn’t only arise with bird cervicals. The same issue arises in measuring more sensible and elegant vertebrae, such as our old friend HMN SII:C8, or MB.R.2181:C8 as we must now learn to call it.
Although the back of the vertebra is nice and simple here — it’s obvious what line we’re measuring to at the back — we have three choices of where the “front” of the vertebra is, and a case can be made for any of them as being “the length of the vertebra”.
The longest measurement (here marked “T” for “total length”) goes to the front of the prezygapophyseal rami. The next one (“C” for “centrum length”) goes to the anteriormost point of the condyle. The distinction is important: as noted recently, the longest vertebra in the world belongs to Sauroposeidon if we use total length, but to Supersaurus if we use centrum length.
But in life, most of the condyle would be buried in the cotyle of the preceding vertebra. So should it count towards the length of the vertebra? If you consider a string of articulated vertebrae, the buried condyles don’t contribute to the overall length of the neck. So Matt and I call the length from the posterior margin of the condyle to the posterior margin of the cotyle the functional length (marked “F” above), which I believe is a new term.
Another way to think of the functional length is the distance from a given point on a vertebra (in this case the posterior margin of the cotyle) to the same point on the adjacent vertebra:
For our current project, Matt and I are interested in how the lengths of individual vertebrae contribute to total neck length, so for our purposes, functional length is definitely what we want.
By the way, Janensch is the only author I know of to have even recognised the importance of functional length. The measurement tables on pages 39 and 44 have columns for “Gesamtlänge des Wirbels ab Vorderende per Präzygapophyse”, “Gesamtlänge der Wirbel-Körpers in 1/2 Höhe” and “Länge der Wirbel-körpers ohne Condylus in 1/2 Höhe” — that is, ”Total length of the vertebra from the anterior end of the prezygapophysis”, “Total length of the centrum measured at mid-height” and “Length of the centrum minus condyle at mid-height”. This is typical of his careful and methodical approach. Kudos!
Hey! I thought this was about turkeys
And so it is. Here is the functional length measurement for a turkey cervical:
It’s the shortest anteroposterior distance between the two articular surfaces.
Measuring functional length
Matt and I chatted about this at some length, and I am ashamed to say that we thought through all sorts of complicated solution involving subtracting measurements from known scaffold length and suchlike.
It took us a stupidly long to to arrive at the very obvious solution, which is just to modify the calipers to have a “tooth” that can protrude into the concavity of the anterior articulation between its left and right lips. Easily done with a flat-ended screw and a blob of wood glue:
With the measurements of all the vertebrae in my series, I can now fairly confidently expect that the sum of the individual lengths will come out at about the length of the complete neck.
You know, unless intervertebral cartilage turns out to be important or something.
- Janensch, Werner. 1950. Die Wirbelsaule von Brachiosaurus brancai. Palaeontographica (Suppl. 7) 3:27-93.
1. Matt and I are so used to opisthocoelous sauropod presacrals that when we’re talking about vertebrae — any vertebrae — we tend to say “condyle” and “cotyle” for the anterior and posterior articular surfaces, no matter what their morphology. When talking about crocodile cervicals or titanosaur caudals, we’re even likely to say ridiculous things like “the condyle is concave and the cotyle is convex”. Nonsense, of course: condyle means “A rounded prominence at the end of a bone, most often for articulation with another bone.” What we should say is “the condyle is at the back and the cotyle is in front”.
January 10, 2013
I’ve measured a few necks in my time, including the neck of a baby giraffe. I can tell you from experience that necks are awkward things to measure, even if they have been conveniently divested of their heads and torsos. They have a tendency to curl up, which impedes attempts to find the straight-line length. Even when you manage to hold them straight, you want them maximally compressed end-to-end rather than stretched out, which is hard to achieve without buckling them out of the straight line. And then you need to measure between perpendiculars in a straight line.
Tonight, I needed to measure the mass and length of seven turkey necks. (Never mind why, all will become clear in time.) And I found a way to do it that works much better than anything I’ve done before.
Here’s the equipment:
You will need:
- Kitchen scales (for weighing the necks)
- Small numbered labels (for the sandwich bags that the necks will go into for the freezer once they’ve been measured)
- Pen and paper to take down the measurements
- Translucent ruler
- Saucepan full of turkey necks
- Slightly less than one half of a birthday cake decorated like a map of Middle-earth [optional]
- A Duplo baseboard (double-sized Lego) and about fifteen 4×2 bricks
Use the bricks to build an L-shaped bracket on the board — about half way back, so that can rest your hand in front of it.
Now you can push the neck into the angle of the bracket. By keeping it pressed firmly against the back wall (yellow in my construction), you can keep it straight. I find the best way to get the neck exactly abutting the left (red) wall is to start with the neck in its natural position, with the anterior and posterior ends curving towards you, then sort of unroll it against the back wall, and finally push the posterior end into place with your little finger (see below). There is a satisfying moment– almost a click — as the back end pops into place and the neck slides along a little to right as necessary to accommodate the added length.
Now use another brick (blue in this photo) as a bracket: slide it along the back wall from right to left until it’s solidly abutting the anteriormost vertebra. If you do this right, there is very little travel: the entire series of vertebrae is lined up and solidly abutted, with bone pushing against the left wall and your new brick. I find there’s less than half a millimeter of variation between the length under gentle-but-firm pressure (which is what I measured) and under the very strongest force you can exert without buckling the neck.
Once you have found the blue brick’s correct position, you need to hold it firmly in place and measure its position relative the the left wall. (It doesn’t matter if you let the neck re-curl at this point, so long as the blue brick doesn’t shift.)
You need a translucent ruler so that you can lay it across the neck and see where blue brick falls under the scale. (My ruler’s zero is, rather annoyingly, 5 mm from the end; so I needed to subtract 5 mm from the lengths I measured.)
Finally, I bagged up each neck in its own sandwich bag, ready for the freezer. Each neck is labelled with a number so that when I take it out for dissection, I will be able to relate the measurements and observations that I make back to these initial measurements.
For the record, here are the measurements:
- Neck 1: 154 g, 179.5 mm.
- Neck 2: 122 g, 151 mm.
- Neck 3: 154 g, 199.5 mm.
- Neck 4: 133 g, 162.5 mm.
- Neck 5: 142 g, 169 mm.
- Neck 6: 80 g, 167 mm.
- Neck 7: 70 g, 169 mm.
As expected, there is some correlation between neck mass and length; but not as much as you might expect. Naively (i.e. assuming isometric similarity) mass should be proportional to length cubed, but there is a lot of scatter about that line. I don’t know whether that is due to individual variation, or merely because the various necks — all of them incomplete — are different sections of the full neck. Hopefully I will be able to confirm or rule out that possibility when I’ve dissected down to naked vertebrae.
November 27, 2012
The best open-access publishers make their articles open from the get-go, and leave them that way forever. (That’s part of what makes them best.) But it’s not unusual to find articles which either start out free to access, then go behind a paywall; or that start out paywalled but are later released; or that live behind a paywall but peek out for a limited period.
Let’s talk about these.
Initial “open access”
You’ll sometimes come across journals where articles are free to read for some initial period after their publication. For example, the announcement of the Journal of Photonics for Energy says “The journal will be available as open access for the first year”; and the 2008/9 progress report for the Journal of Nutrition says “We will continue to restrict open access for one year, as per current procedure”.
Despite the good intentions of the journals, these articles are not open access in any useful sense. The point of an open-access article is that it’s there when you need it. If it’s there this week, but I need to read and cite it next week when I can’t get it any more, then that’s no good.
Of course, this doesn’t mean that publishers have a mandate to keep articles up on their web-sites forever (although we would prefer that they do). What it means is that, if they want to be open access, they can’t prohibit others from mirroring and archiving those papers, and continuing to make them available after they’ve disappeared from the publisher’s site.
Note that any article published with a Creative Commons licence — even the most restrictive of those licences — is safe from this kind of disappearance. Those licences guarantee third parties’ rights to archive, replicate and redistribute the articles.
Delayed open access
it’s probably more common to take the opposite approach. Some journals, including Science and Proceedings B, make articles free to read, and so “gratis open access”, after an embargo period during which they are available only to subscribers. This period is one year in the case of both these journals; that seems to be typical.
Are such journals open access? I would say that the journals themselves are not open access, but that the articles become open access once they cross the release line. So for example, Raichlen and Polk’s new neurobiology paper in Proc B. is not open access, but Anderson et al.’s seed-dispersal paper (which is a year older) is. On that basis you might choose to refer to Proc B. as a “delayed open access” journal.
[Unfortunately, Science is not truly open access even for older articles such as Stevens and Parrish's DinoMorph. That's because it requires registration/login before you can get to the papers. The BOAI FAQ does not accept registration-required content as open access, specifying "without financial, legal, or technical barriers other than those inseparable from gaining access to the internet itself".]
Transitory “open access”
And then you have the worst of both worlds. Every now an then a journal or a publisher has a special offer where they open up access to their articles for a limited period — for example, this one where the Royal Society opened up all their content for two weeks, or this where an issue of European Physical Journal D was opened for a week.
It seems churlish to criticise a generous action like this, but I find it close to useless, and I think most other researchers will, too. When I am working on a paper, I don’t choose what to cite based on which journal or publisher the papers are from: I would never think, “Oh, let’s see, European Physical Journal D is open at the moment, I’ll cite something from there”. I cite what’s relevant and appropriate, irrespective of its source; and if I can’t get the papers I need at that time there’s a problem.
I sometimes wonder what publishers think will be the result of this kind of limited-time-only offer. One obvious outcome is that people will batch-download the transitorily available content — either to store up for themselves in case they even happen to need it (which is wasteful of both bandwidth and storage); or to post openly elsewhere for permanence (which is usually a violation of copyright).
To summarise: I think that making articles open access after a delay is a good thing (though obviously not as good as making them open access immediately!). But that making them free to read for a limited time — either when first published or as part of some special event later — is of very limited value, and can’t really be described as open access.
In the previous section, we discussed the various licences that can be used for open-access articles. But that may have been premature, because licences are agreements whereby copyright holders waive some of their rights, and we hadn’t actually talked about copyright first. So let’s do that now.
(This post is relevant to subscription publishing as well as open access.)
Who owns copyright in a new work?
In general, whenever you do any creative work, you own the copyright of that work — whether it’s painting a picture, composing a tune or writing an academic paper. You don’t need to do anything special to obtain copyright (although back in the day registration used to be necessary in some jurisdictions).
There are a couple of exceptions to this general rule. One of the most important is a work for hire. When you are paid to create a copyrightable work, the terms of the contract under which you do it may stipulate that your employer owns the work. This is common, and quite reasonable, in some situations: for example, my day-job employer Index Data owns the copyright to the code I write for them. It’s much less common in other situations — in particular, I have never heard of a university taking copyright for its researchers’ works.
Another important exception to the usual you-wrote-it-you-own rule is that all work created by the US Government is public domain: that is, no-one owns it, there is no copyright, and anyone can do whatever they like with it.
By the way, anything that you own the copyright for, including your manuscripts, you can place in the public domain. (The Creative Commons CC0 tool exists to help you do this unambiguously.)
When your manuscript is accepted by a journal for publication, many publishers will ask you to transfer copyright to them, often insisting that it is an absolute requirement for publication. They have a form for you to sign; once you have done this, you no longer own your work. (Since this post is supposed to be purely expository and not at all evangelistic, I will refrain from comment on this issue.)
There are a few alternatives to copyright transfer.
First, if the work is already in the public domain — for example because it was created by an employee of the US Government as part of their job — then there is no copyright to transfer. In this case, the journal is already free to publish the work (as indeed is any other journal), and they will usually just ask the author to make a statement certifying the PD status of the work. (JVP’s form has this option.) I don’t know how such journals would react to non-Government employees who had dedicated their work to the public domain. It would be interesting to find out.
Second, some journals do not require the author to transfer copyright, but to give them a licence to publish. Although this can be a liberal option, such licences often impose many restrictions on the author, so that you’re not allowed (because of having signed the contract) to do things that you might, as copyright holder, expect to do. In some cases, the value of having retained copyright is very small.
Finally, most open-access journals that use a true open-access licence such as CC BY allow authors to retain copyright because, well, it makes no difference to them who holds it. Whoever the copyright holder is, third parties are in the exact same situation: they are free to do whatever they want with it provided that they acknowledge the authorship.
Note that authorship is a different matter from holding copyright. Even if you transfer copyright of your work to a publisher, you still have the right to be identified as the author (unless of course you waive that separately).
How licences work with copyright
If you hold copyright in a work, I have essentially no rights regarding it, beyond what you give me. I can’t print it, I can’t make copies for myself or others, I can’t post it on my web-site, I can’t translate it, I can’t make derived works, I can’t use images from it on Wikipedia, and so on.
So in that state, it’s essentially useless.
This is why you release your work under a licence. Licences waive some of the copyright holder’s rights (i.e. some of the restrictions placed on others’ use of your work). You may do this for financial gain (“you can sell copies of my book provided you pay me 10% of the cover price as a royalty”), or out of the goodness of your heart (“public money paid me to do this research, so I am giving the result to the world.”)
The specific set of rights that you give varies with the license. Restrictive licences may give the right only to read the work. Others may allow various forms of re-use, perhaps limited as to extent or field of applicability. Some licences allow certain forms of use in non-commercial contexts, and less in commercial contexts. Choosing an appropriate licence is an art-form: we’ll talk more about it in a subsequent post.
Only the copyright holder can grant someone a licence to their work. The holder may choose to give different licences to different people. For example, you might make your paper available under CC BY-NC, which forbids commercial use; then if someone wants to use your work commercially, they might offer you money in exchange for furnishing the work to them on terms that allow commercial use.
Epilogue: a note on patents and trademarks
Patents, copyrights and trademarks are often referred to under the blanket term “intellectual property”, or IP for short. This is unfortunate: they are actually three very different branches of law, and have little in common. That’s one reason that I avoid the misleading term “intellectual property”.
Patents cover inventions, or increasingly often ideas. They must be applied for and paid for, and have a strictly limited term — 20 or 14 years in the US — after which they expire and anyone can use the invention.
Copyrights cover specific works, which may use inventions but differ from them in having a specific fixed form. They apply automatically and don’t need to be applied for; and they last much longer than patents. (The details of copyright terms are complicated, but generally anything you create now will remain under copyright in the US until 70 years after your death.)
Trademarks cover names and logos and exist to prevent consumer confusion that could arise if two similarly named organisations or products exist in the same space. They must be applied for.
Here are some examples. I might have patented the idea of using CT scanning to look inside sauropod vertebrae, if I’d thought of it first. (I’d have had to write the description in terms that sound like an invention, an actual device for using CT.) I couldn’t copyright or trademark that idea. I would hold copyright in any specific paper I wrote using that technique, but could not patent or trademark that paper. And if I started a company to do CT scanning of sauropod vertebrae, I could register its name as a trademark, but I couldn’t patent or copyright it.
For the purposes of academic publishing, copyrights are by far the most important of these areas; patents can arise from research, and introduce their own issues, which we won’t discuss here; trademarks are irrelevant.
Thanks for sticking with this series. In part 1, we looked at what open access means, and what terms to use in describing it. In part 2, we considered the Gold and Green roads to open access. In part 3, we touched on zero-cost Gold OA, sometimes known as “Platinum”. This time, we’re going to get down the nitty gritty of the actual licences that govern what you can do with a paper that you’ve downloaded.
As usual in this series, I will try to keep my opinions and preferences out of it, and limit myself to uncontroversial statements. So for example, I will not express a preference for one Creative Commons licence over another, even though I do have a preference.
Unfortunately, this is still very common. Lots of journals that make their articles freely available to read online say nothing about what you are and are not allowed to do with them. PalArch’s journal of vertebrate palaeontology is one of these — I have no idea, for example, whether I am allowed to print a copy of an article for myself; or, if I am, then whether I can give it to a friend; or if I can print three copies for three friends, or fifty copies for a group of students.
Not much better is the sort of vague statement given by Palaeontologia Electronica:
All articles appearing in Palaeontologia Electronica (PE) are available free of charge from the World Wide Web through the Palaeontologia Electronica Site. Copyrights for technical articles (text and graphics) are assigned to Palaeontologia Electronica Sponsors where appropriate … If you would like to distribute copies of materials published by Palaeontologia Electronica we encourage you to obtain the requisite permissions from the copyright holders.
The implication here is that I can print a copy for myself but not for my friend, but it’s not at all explicit. (Let’s leave aside for the moment there there’s any reason for such a condition, and limit our questions to what the conditions are, not what they should be.)
So the first thing to say about open-access licences is: please have a licence. Even if it’s a horrible, restrictive licence, please at least be clear about it. Merely shoving PDFs up on the web and walking away is asking for misunderstanding.
One step up from no licence at all is a custom licence, written for a particular journal or publisher. One such is the set of terms used by Elsevier for their “sponsored articles”. (Credit to Elsevier for making these fairly easy to find now — it was not always the case!)
Leaving aside how restrictive these terms are, let’s at least give credit where it’s due, and acknowledge that they are explicit. The problem is, it’s a lot to read and understand. Elsevier’s terms are actually fairly short and sweet as these things go: 300-odd words. But it’s not unusual for these things to be multi-page monsters. Who can read and understand the implications of such things? If only there were a small set of simple, well-defined standard licences, so that content providers could just pick the one they wanted and everyone would know what it meant.
Creative Commons licences
… and that is the purpose of Creative Commons. There are about seven different Creative Commons licences, depending on how you count them, but they are made up from a small number of easy-to-understand building blocks. Since each such block has a two-letter name, it’s easy to name a specific Creative Commons licence such as BY-NC-SA. (The full abbreviations of the licences begin with “CC”.)
CC BY is the basic CC licence, and say that you are allowed to do anything at all with the content of the article provided only that you credit the author. It’s the licence used by the biggest and most influential open-access publishers (PLOS, BMC, Hindawi) precisely because it allows the licenced work to have the most value. Wikipedia uses it for the same reason (as indeed does this blog). When dealing with a CC BY article, you can reuse passages of it in your own work, copy its illustrations into a Wikipedia article, hand out copies to classes you teach, extract numeric data and add it to your database, and so on.
You can augment — or, rather, restrict — the CC BY terms by adding other clauses:
The NC clause means “non-commercial”, and restricts downstream use of the work to non-commercial contexts — although exactly what that means is vague and difficult to define. The purpose of this clause is to ensure that if anyone makes money from the work, the author gets a slice. (We’ll discuss this more in a future post.)
The ND and SA clauses are two alternatives: you can’t have both together, that would be a contradiction:
ND means “no derivatives”: you’re allowed to make copies of the entire article, but not to “remix” it: you can’t make translations, extract passages, adapt it into a blog post, etc. The idea of this clause is to protect authorial integrity.
By contrast, SA means “share alike”: you are allowed to make derivatives, but only on the condition that you release them under the same licence. The idea here is to make openness viral, to ensure that it’s passed on to other projects.
Finally, there’s CC0. This is not exactly a licence, but a formal declaration that the work is placed in the public domain, that copyright is waived, and that can you do whatever you like with it, subject to no conditions at all, not even attribution. (Some other classes of work are also in the public domain, notably anything produced by US Federal employees, including those who work for the BLM.)
These various CC licensing options can be stacked to make the following licences: CC BY, CC BY-NC, CC BY-ND, CC BY-SA, CC BY-NC-ND and CC BY-NC-SA. And CC0 makes seven.
SIDEBAR: If you’re familiar with the major open-source software licences, you’ll recognise CC BY as being similar to the Apache and BSD licences; and CC BY-SA as similar to the GNU General Public Licence (GPL). There are no open-source software equivalents of CC licences with the NC or ND clauses, as these would violate the open-source definition. CC0 is of course equivalent to public domain software.
Note that, as with any other licence, you have the option of routing around CC licences by negotiating with the copyright holder — which is often, though not always, the author. If for some reason you particularly wanted to reproduce an SV-POW! article and not credit me as the author, then this blog’s CC BY licence doesn’t give you permission to do that — but you can contact me and ask whether I’ll allow it anyway. More realistically, if you wanted to use CC BY-NC material in your business’s training materials, you might be able to negotiate its use, for a fee.
No doubt there are other licences out there other than the CC ones and the ones that various publishers make up for themselves. (In the software world there are lots of these, to no-one’s benefit.) But I can’t think of any examples. Can anyone?
One last nasty problem needs to be mentioned. While journals tend to at least be consistent in the terms under which they make articles available, repositories often are not. For example, articles in arXiv are provided under four different conditions: CC BY, CC BY-NC-SA, public domain, and an underspecified “licence to distribute“. Worse still, I can’t see that their pages even specify which licence a given article uses.
This makes it harder, in general, to safely reuse content from repositories. It’s one reason why some people favour Gold OA over Green OA.
November 18, 2012
As we saw last time, the appeal of the Gold route to open access is that the publisher does the work of making the article freely available in an obvious, well-known place in its final typeset format. Conversely the appeal of the Green route is that it doesn’t cost the author or her institution any money.
What happens when we combine these two advantages, and get publishers to typeset, publish and archive open-access articles at no charge?
Yes, it does happen. One outstanding journal that does this is Acta Palaeontologia Polonica — this is one of the reasons that I have published there twice (neck posture, Brontomerus) and Matt has three other APP papers as well as being co-author on those two. Another is Palaeontologia Electronica. In a different field, the Journal of Machine Learning Research (JMLR) is of particular interest because the article An Efficient Journal explains in some detail how it’s done.
According to the definitions I gave last time, this best-of-both-worlds scenario is in fact the best case of the Gold route: because the key point is that the publisher is responsible for making the work freely available. The process of publication in these venues is identical to that in other Gold OA venues — the only difference is the lack of a fee.
Recently, I’ve started to hear two new terms used to describe zero-fee Gold open access: Platinum OA and Diamond OA. I am not very keen on either, because they give the incorrect impression that there is another route to open access, fundamentally different from either Gold or Green. But if either term is to be used, “Platinum OA” is a better term than “Diamond OA” because at least platinum is a precious metal like gold — so the connotation is “like Gold OA but ever better”.
So I recommend not using the term “Diamond OA”.
Last time, we looked at what the term “open access” actually means. We noted that its been widely abused, so that when you need to be specific about the full meaning you need to say “BOAI-compliant”; we recognised that much of what is described as OA is really only “gratis OA”, or as Ross Mounce called it, “gratis access”; and we noted that the term “libre open access” is literally meaningless and should be avoided.
At the moment, the big argument within the open-access movement is about Gold vs. Green open access. This time, we’ll look at what these terms mean, what they don’t mean, and some of the pros and cons of each.
What Gold and Green mean
Gold open access means that that publisher, which creates the final published form of the paper (i.e. usually a PDF) makes that final published form freely available.
Mostly that means they will host it on their own site, as for example BMC and PLOS (two Gold-OA publishers) do. In some cases, the papers may be hosted off-site: for example, eLife doesn’t host its own papers at the time of writing, but leaves them for PubMed Central to host. The key point is that the publisher is responsible for making the work freely available.
Well-behaved Gold-OA publishers will also do things like ensuring the papers are indexed in reference databases like PubMed, and that they are archived in schemes like LOCKSS and Portico.
Green open access means that the publisher locks the final published form of the paper behind a paywall, but the author takes steps to ensure that it’s freely avalaible elsewhere.
The form of the paper that is made available varies: ideally it’s the final published form; sometimes it may be the final accepted manuscript, as it was when the author last touched it, before the publisher typeset it.
Often, Green OA uses institutional repositories (IRs). Another common option is a subject repository, of which the best known is arXiv — the vast preprint archive for maths, physics and astronomy, and occasionally palaeontology. Another (rather weak) form of Green OA is individual researcher collections on web-pages, such as Matt’s and mine. There may be other options, such as uploading the manuscript into a torrent space, and letting the world mirror it. The key point is that the author has to make this happen, rather than leaving it to the publisher.
Gold and Green are strategies
The first and most important thing to understand is that Gold OA and Green OA are not two different goals. They are two complementary strategies for reaching the same goal — which is open access.
A given downloaded paper is not a Gold OA paper, or a Green OA paper. It’s just an open access paper. It’s true, of course, that it reached its user by means of the Gold or Green routes. But by the time it’s arrived at its destination, the route it took is no more interesting than whether I took the M40 or M4 on my journey to London. Being in London is what matters.
This is why the loathing that some Green advocates seem to feel for Gold is so misplaced. We want to get to London. I may find the M40 route more convenient, but I really don’t want to get into a situation where I’m insulting those who chose to drive down the M4.
“Gold” does not mean “higher quality” or “more open”
In retrospect, it may be a bit of a shame that the Gold strategy was given the name Gold, which connotes quality, rather than a more value-neutral colour such as blue. You sometimes read people writing about Gold OA as though it’s the gold standard — the best OA you can get! But of course that’s not true.
The best OA you can get is OA that complies to the original definition of the Budapest Open Access Initiative (BOAI): that is, open access that permits all kinds of reuse as well as merely reading. You can achieve BOAI-compliant OA by the Gold or Green routes. And you can also take the Gold or Green routes to a paper that is merely “gratis access” — i.e. free to read, but with all other rights reserved. Whether the Gold or Green route was taken tells you nothing about what rights you have.
Advantages and disadvantages of Gold and Green
So why does it make any difference whether Gold or Green is used? Well, there are a few things:
- Green requires more work from the author
- Gold may require the author (or, more usually, her institution) to pay a publication fee.
- Green may undermine publishers’ businesses. [Whether that is an advantage or disadvantage may be open to argument.]
- In the common case where Green provides only the author’s final manuscript, there are two versions of the paper out there, which can cause confusion.
- In that case, you don’t generally know the final published version’s page-numbers when working from the author’s manuscript, which means you can’t cite pages.
- Fragmentation of papers across many Green repositories causes problems:
- It can be hard to find a paper — there is no good cross-IR aggregator, and who wants to find the lead author’s institution’s IR on the way to discovering the paper?
- Different IRs impose different bizarre and unnecessary reuse conditions, often preventing BOAI compliance.
- Coverage of IRs is surprisingly patchy, and some very well-respected universities doesn’t seem to have one at all.
So the equation is a fairly complex one, and it’s perhaps not too surprising that some OA advocates prefer Gold on balance, while others feel that Green is better. (I may have my own opinions on this issue, but I’ll leave them out of this post.)
Green and Gold in open-access mandates
The best-known and most influential open-access mandate is that of the National Institutes of Health (NIH) in the USA, which requires that authors “Submit papers to PubMed Central (PMC) and approve public release”. This is is of course a form of Green OA, with a particular repository wired in.
The great majority of subsequent open-access mandates, whether from funding bodies or universities, have followed the NIH’s lead in requiring Green OA. Why? I don’t really know — mandating organisations don’t tend to discuss their reasoning. But one obviously appealing aspect of the Green route is that no-one has to think about money. In particular, funders don’t have to find more of it to pay for publication.
Against this backdrop it’s been encouraging to see that many of the more recent mandates are neutral on which route should be taken, caring only that open access is achieved. An obvious example is the new RCUK policy, which describes both Gold and Green (though without using either term).
[It's true that RCUK has expressed a preference for Gold (and is providing money to make it possible), but they are clear that the choice is one for authors to make. The specific reasons given for preferring Gold in the linked post seem spurious, as I noted in my comment there, but that doesn't mean there aren't legitimate reasons.]
The Finch report strongly favours Gold
Finally a word on the Finch report, recently written to guide UK open-access policy. This report was produced by a committee containing researchers, librarians, administrators, and — crucially — publishers. Now publishers hate Green OA, because it doesn’t generate revenue for them. And it seems that the publisher lobby nobbled the otherwise excellent report by excising all mention of Green OA. My feeling is that expressing a preference for Gold would have been reasonable, but that pretending Green doesn’t exist was misleading and irresponsible.
The very unfortunate consequence of this has been that certain open access advocates who strongly favour the Green route have become very noisy with articles and comments that give the impression they’re opposed to open access — because the Gold emphasis of the Finch report (and the mild Gold preference of the RCUK policy) is different from the way they would like to do things. Folks, please stop this. It doesn’t help anyone, with the possible exception of Elsevier.
The real point here is that the world needs open access. How it gets open access is a very secondary issue in comparison.
November 15, 2012
I’m going to keep this free of advocacy. Hopefully everything I say here will be uncontroversial, because all I am doing is surveying definitions and clarifying distinctions. I’ll save my opinions for later articles (not that there is any secret about them).
Open access (or OA)
It may seem a bit surprising to have to define “open access” when we’ve all been talking about about constantly for a year. But one of the biggest issues that derails constructive discussion of the move to OA is that the term is used in different ways by different groups.
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.”
Note, then that, open access as originally defined means much more than the ability to freely read a paper on the Internet, but also allows much wider range of activities — including all forms of redistribution, repurposing and content-mining.
Of course it’s also a useful thing when an article is made available with fewer rights — typically only the right to view an article on the Internet, or sometimes to download a personal copy or make a printout. The right to do these things is valuable, and needs a name. Unfortunately, the name it’s often been given is “open access”, obscuring the important distinction between true open access and the more restrictive form of access-to-read.
This confusion is a very bad thing. Why? Well, first of all, because we’re scientists and we care about what terms mean. If I say “dinosaur” you know that I mean “descendant of the most recent common ancestor of Megalosaurus and Iguanodon“, and that I am not going to suddenly start talking as though Dimetrodon or Sarcosuchus were a dinosaur. In just the same way, if I say “open access”, you need to know what I’m talking about.
Second, there are big practical implications. The one that’s in the public eye at the moment is text-mining — having computers, rather than humans, read papers and extract the raw facts from them. When articles are open access in the original sense of the term, there is no ambiguity about whether you’re allowed to text-mine them or not. When they are “open access” only in the more limited sense, you may or may not have the right you want. Often you can’t even tell, and the best you can hope for is that you’ll emerge on the other side of long, complex negotiations with permission. True open access has other important implications, such as availability to be incorporated into Wikipedia.
BOAI, BBB, “full open access”, @ccess
To be more explicit about what particular rights are readers are given regarding an open-access article, several terms have been used.
- The one I like best is BOAI-compliant, which refers explicitly back to the Budapest initiative (and has the advantage of using an acronym that can’t be confused with anything else).
- You might also see BBB, which stands for Budapest/Berlin/Bethesda, the names of three very similar open-access declarations.
- You occasionally even see the redundant BOAI/BBB. Please don’t.
- Phrases such as “full open access” and “true open access” are sometimes used, but they don’t help because they are just as prone to abuse as unadorned “open access”.
- Finally, the @ccess group started to use the term @ccess to refer to BOAI compliance. Although I am part of the @ccess group, I don’t think that adding yet another term has helped (and I am pleased to see that they seem to have dropped that usage).
So the unfortunate consequence of the unfortunate broadening of the meaning of “open access” has been the coining of four or five different terms all intended to indicate what was originally meant by the original term. Very unhelpful.
Gratis vs. libre
The most deliberate attempt to clarify exactly what degree of freedom is meant by “open access” was in a 2008 memo by Peter Suber, with input from Stevan Harnad. It defined the terms “gratis OS” and “libre OS”, to mean “free as in beer” and “free as in speech” respectively. That is, “gratis OA” means only the removal of price barriers, and says nothing about permissions.
So Elsevier’s “sponsored articles” could be described as “gratis OA”.
Unfortunately (and I seem to be using that word a lot!) the usually reliable Peter Suber completely fumbled the ball in the definition of “libre OA”:
I’ve decided to use the term “gratis OA” for the removal of price barriers alone and “libre OA” for the removal of price and at least some permission barriers. [...] There is more than one kind of permission barrier to remove. Therefore, there is more than one kind or degree of libre OA.
This means that the term “libre OA” is completely useless. It tells you literally nothing about what you can do. Can you republish a libre OA article? Can you text-mine it? Can you use figures from it in your own work? You can’t tell. You may be able to do any or all of these; but the fact that the article is libre OA doesn’t tell you that.
A publisher could make an article free-to-read and add the stipulation that you’re allowed to reuse portions of it in your own work provided that your work is printed on sheets of pure diamond using ink made from snow-leopard foreskins. And that wholly useless concession towards reuse would suffice to make the article “libre OA” under the Suber/Harnar definition.
Conclusion: recommendations on terminology
- Use “open access” (or OA) when talking in general terms, but be aware that in practice its meaning is vaguer than when originally defined
- When more precision is needed, use “BOAI-compliant” to mean open access as originally defined. Avoid “BBB”, “full open access” and other such alternatives.
- It’s fine to use “gratis OA” to convey that something is free to read but offers no further permissions. But:
- Do not use “libre OA” because it’s meaningless.
In short, the only terms you need to use are “open access”, “BOAI-compliant” and maybe “gratis OA”.