In accordance with Björn's commitment to openness in science, he circulated a preprint and paid to make the published version Open Access in the hope of stimulating further discussion:

The article has been through several rounds of peer-review, both informal and formal [...] since august this year. Of course, the real discussion, I would hope, isn't starting until today, when the article actually became accessible.

I read the preprint, and it made my head ache. In a good way. I'm really not qualified to say whether Björn is right or wrong or completely nuts on this issue, but he's taken an ambitious stab at a Big Question and that's always good. More to the point, he's done it well and carefully and it's worth your time to play along at home.

Do your brain a favor and give it a workout -- the full article is freely available online, and if you have substantive comments to make I guarantee you that the author will be delighted.

To whet your appetite, here's the abstract:

Until the advent of modern neuroscience, free will used to be a theological and a metaphysical concept, debated with little reference to brain function. Today, with ever increasing understanding of neurons, circuits and cognition, this concept has become outdated and any metaphysical account of free will is rightfully rejected. The consequence is not, however, that we become mindless automata responding predictably to external stimuli. On the contrary, accumulating evidence also from brains much smaller than ours points towards a general organization of brain function that incorporates flexible decision-making on the basis of complex computations negotiating internal and external processing. The adaptive value of such an organization consists of being unpredictable for competitors, prey or predators, as well as being able to explore the hidden resource deterministic automats would never find. At the same time, this organization allows all animals to respond efficiently with tried-and-tested behaviours to predictable and reliable stimuli. As has been the case so many times in the history of neuroscience, invertebrate model systems are spearheading these research efforts. This comparatively recent evidence indicates that one common ability of most if not all brains is to choose among different behavioural options even in the absence of differences in the environment and perform genuinely novel acts. Therefore, it seems a reasonable effort for any neurobiologist to join and support a rather illustrious list of scholars who are trying to wrestle the term 'free will' from its metaphysical ancestry. The goal is to arrive at a scientific concept of free will, starting from these recently discovered processes with a strong emphasis on the neurobiological mechanisms underlying them.

XXXI

WITH rue my heart is laden
  For golden friends I had,
For many a rose-lipt maiden
  And many a lightfoot lad.

By brooks too broad for leaping
  The lightfoot boys are laid;
The rose-lipt girls are sleeping
  In fields where roses fade.

--A.E. Housman

This is why we can't have nice things.

Fortunately, my friend Abel has done a great job of introducing and explaining the day and its background, so I'll just point you there: What is this Juneteenth of which you speak?

The FUD merchants want to claim that, if no citation advantage exists, there is no point to Open Access: that unless OA papers are currently garnering more citations than their TA equivalents, current levels of access must be adequate; or that if OA papers, which presumably are read more, are not cited more, then OA must be a repository for the second rate. Hence the controversy: it's an easy way to obscure the debate, sending up a cloud of statistical argument like a fleeing cuttlefish squirting ink. "Look over there, OA proponents are wrong about this, surely they must be wrong about everything, pay no attention to the massive profits behind the curtain."

Considering that:

• only around 20% of the world's scholarly journals are OA, and very few of those have been around for more than about a decade
• the subscription model is propped up by systemic inertia and by the fact that it simultaneously makes competitors into complements and hides its costs in multiple places, thereby divorcing demand from the reality check that costs ought to provide; and
• although we might wish it were otherwise, the practice of citing papers known only from their free-access abstracts or mention in another paper is hardly uncommon

it's something of a miracle if any OA citation advantage shows up anywhere. More importantly, though, the citation advantage was always a minor point in the list of reasons to prefer Open to Toll Access:

(1) Not everyone who needs to read the primary literature is going to write anything citing it. That doesn't make providing them with access to the literature any less important, and no payment or institutional affiliation is required to read Open Access information.

(2) Toll Access confines data- and textmining to isolated, artificial commercial sections of the body of knowledge, hindering progress on mining methodologies, restricting the reach of existing work and precluding any idea of a comprehensive protocol.

(3) OA provides better value for money than Toll Access. Regardless of where the money comes from, OA is a one-time up-front expense that covers all subsequent use: pay the midwife, but keep the baby. Peter Suber has written a careful exposition of this argument from the taxpayer perspective, but most if not all of his points map readily onto any research funder.

(4) Open Access scales where Toll Access doesn't; my own recent estimate (caveat lector!) is that library access, even at the best funded libraries, runs to around half of the total available scholarly journal literature. What use is a system that enables publication without enabling access?The subscription model divorces (part of) the cost of dissemination from the overall cost of production of scholarly information, which has allowed research funders to overlook that part of the cost of their mission. It's been historically picked up by libraries, but that's easily revealed as a shell game when you look at where library funding comes from. Who loses the shell game? Academics whose work is less widely available than it should be, and anyone who wants to read the primary literature. Who wins? Publishers, whose prices have been allowed to escalate because they have largely escaped scrutiny (except by librarians, who for no good reason that I can see have been largely ignored, at least until relatively recently, by academic and political decision makers).

So I don't think the citation advantage question has much bearing on the larger question of the value of Open Access. Existing access is clearly inadequate by comparison with Open Access by way of points (1) and (2) above, and OA is in any case preferable by way of point (3) and inevitable by way of point (4).

Two well scrubbed boys
in suits and ties
ruined my morning with their talk,
feigning interest in my book
and where I'm from
and friendly lies
the way they're taught, to draw me in
till they can teach me of my sin.

Listen here, my shiny lads,
my would-be rescuers of soul—
but here's your stop, so shake my hand
and go your way and wish me well.
A narrow Heaven chafes your thought,
quite unlike my light, loose-fitting Hell.

• In 2003, Kathy Varjabedian at LANL compared the electronic holdings at 12 (large, well funded research) libraries with the ISI Journal Citation Report's top 100 most-cited journals for the previous year, producing an estimate for the ullage of those libraries of between 2% and a startling 54% (or 0% and 40%, if clinical titles were excluded).
• Also in 2003, Carol Tenopir estimated that there were around 44,000 scholarly journals in publication, just over 21,000 of them "refereed", which is the best proxy that Ulrich's Periodicals Directory allows for "peer-reviewed". Repeating Tenopir's search just now returned 26,677 active, refereed, academic/scholarly journals.
• Last year, I used a UCOSC dataset from 2004, a curated list of about 3000 titles, to estimate the average subscription price for a peer-reviewed scholarly journal (table 2 here) at $1238/title.

• Here are some more data from the Library Journal Periodicals Price Survey:
  
  
  
  Sorry about the jpg, I still can't make MT cope with tables. The spreadsheet is here. In case the image goes awry: the dataset covers more than 5,000 titles from 30 disciplines, and mean price/title is $723 in 2003 and $791 in 2004.
• The mean serials expenditure for an ARL member institution was around $5.5 million in 2003 and $5.8 million in 2004.

At $1200/journal, $5.8 million¹ would buy subscription access to about 4,800 titles, which is less than 23% of the number of active, refereed, academic/scholarly journals. At $700/journal, ARL members -- some of the largest and best funded libraries in America (indeed, in the world) -- are able to afford access to less than half of the scholarly literature.

This seems reasonably consistent with the earlier LANL estimate, given that Varjabedian looked only at the top 100 most-cited journals, which must surely be at the top of any research library's "must-have" list.

It's important to point out that what I'm estimating here is not ullage sensu Suber, but rather library holdings relative to all possible holdings. But I would argue that the access needs of all the scholars and other patrons served by ARL libraries is surely a decent proxy for "all possible journals", if not a significantly larger body of information! Put another way, here I am estimating the gap between current access levels and the information availability of a 100% Open Access world.




-------------
¹This calculation assumes that 100% of the serials budget goes to scholarly journals. That's not true, but I've argued elsewhere that it's at least 90%.

This is Pawel Szczesny at TED Warsaw, describing for the lay public what Open Science is, and what it can mean. Pawel's is the interesting mind to which I refer in the title. I finally met him in person at Science Online earlier this year, but I have been following him around online for years. He never fails to come at a question or problem from an interesting and useful angle, and his TED talk is just the latest example.

What if?

What if I explain in simple words my research area? What if I point you to additional information so you could learn more and understand the topic I am working on? What if I make sure you have access to all relevant literature for free? What if I make sure you have access to all the data so you can play with it on your own? What if I take off this laboratory coat, so there is one artificial difference less between you and me? What if the only thing that mattered in this game of solving nature's mysteries was skills, knowledge and passion? We have a name for that utopian vision: it's called Open Science.

Do yourself a favour, watch the whole thing.

Open science is an iterated prisoner's dilemma, which is a messy and unpredictable business. Too unpredictable for most people to try to build a career on. Thinking about strategies which are likely to be successful leads me towards the concept of an open science community rather than unilateral complete openness - a long term multiplayer collaboration. Does such a community already exist? If not, how do we build one?

It seems clear, though, that such top-down decisions are more likely to be made in academia, and perhaps the nonprofit sector, than in profit-driven industry -- at least until there are enough concrete examples of success to tip the perceived balance of risk. If I'm -- if we Open Foo types are -- right, it's actually riskier to compete than to cooperate in the long term. Better to own a share of X sooner than to delay any return on your investment in the hope of owning X outright later. This is especially true when the resources required to try to own X could be used to get you shares in multiple other projects at the same time.

Even then, openness in industry seems to me unlikely to go beyond consortia. Complete openness (open notebook science) precludes patent protection, and in the dog-eat-dog world of business driven by the insatiable demands of disconnected shareholders, I don't think we are ever going to wean the beancounters off their patents. (We could improve the situation by overhauling the patent process so that teeny incremental changes were not granted full protection, of course; but I digress, and don't get me started.)

So to return to Alan's analogy, "multiplayer" means different things in academia (and perhaps the nonprofit sector) and in business. In business, it means defined communities of co-operation; in academia, I see no good reason why it shouldn't mean everyone (except, perhaps, where the two intersect and academics enter a business-defined collaboration¹).

In academia, communities with an open science focus are beginning to form. The best example is still the one which continues to coalesce around Jean-Claude Bradley's UsefulChem initiative, but it's no longer the only one as it was just a few years ago. Chemist Mat Todd has funding for an open science project to improve synthesis of the anti-schistosomiasis drug, praziquantel. Biophysicist Steve Koch has a labful of open science enthusiast grad students. And so on; there's a list of Open Notebook practitioners on wikipedia, and my own feeling is that technical rather than philosophical barriers are keeping quite a few labs from that list. By being discoverable on the public web, all of these labs can do what Jean-Claude is doing: accumulate collaborators and get more work done. Try searching Google for "DNA tweezers kinesin" -- the second and fifth hits will hook you up with Steve Koch. "Praziquantel synthesis" -- the third hit will take you to the schisto community on The Synaptic Leap, where you'll soon meet Mat Todd, and the seventh hit will take you to a brief discussion of Mat's project on the UsefulChem blog. "Antimalarial Ugi" -- most of the first ten hits will introduce you to UsefulChem. If you're doing something that's in any way related to the work that goes on in these labs, you're one Google search away from a collaboration.

In business, too, more and more companies are recognizing the benefits of wider sharing. Details of private collaborations are hard to come by, but just try searching for "precompetitive sharing" -- even Big Pharma can see that they stand to make net gains from sharing their datasets. For an even better example, check out Sage Bionetworks. I was lucky enough to hear Stephen Friend speak at the Science Commons Symposium a couple of weeks ago, and one of the points he made was that the really big questions in biology require such immense amounts of data that the only way to collect them is to do it in the open. Any impediment at all, be it CC-BY attribution requirements or IP protections, will derail the whole process; the only answer in the end is the public domain.

So, the seeds are there. I think continued crystallization is inevitable, but it's certainly worthwhile to try to monitor and direct the process -- by way of questions like those Alan is asking.

-------------
¹I don't buy the argument, by the way, that unless academics work in secret and enable strong patent protection they will never get industry partners. If you invent something from which profit can be made, someone will want to make that profit. If, without outright patent ownership, it's not enough money to tempt a Roche or an Intel, there will always be smaller, hungrier companies.

I don't know that I buy the idea, but I do notice that my cell phone camera doesn't have a very broad color or contrast palette, so it tends to blow out highlights and lose shadow detail -- and that I'm starting to recognize opportunities to exploit those weaknesses:

Formally, we recommend adopting and acting on the following principles:

1. When publishing data make an explicit and robust statement of your wishes
2. Use a recognized waiver or license that is appropriate for data.
3. If you want your data to be effectively used and added to by others it should be open as defined by the Open Knowledge/Data Definition – in particular non-commercial and other restrictive clauses should not be used.
4. Explicit dedication of data underlying published science into the public domain via PDDL or CCZero is strongly recommended and ensures compliance with both the Science Commons Protocol for Implementing Open Access Data and the Open Knowledge/Data Definition.

I've written elsewhere about my feeling that Open Data/Open Science will eventually need a set of core Declarations to do for the wider movement what the BBB definitions have done for Open Access. A set of widely accepted terms and definitions provides a framework within which ongoing discussions can be much more efficient, focused and useful, as well as a point of reference and a standard introduction for newcomers to a field. Kudos to OKF and partners for making a strong start in this direction.

I do have one small quibble. Following Peters Suber and Murray-Rust, I want Open licenses to be three things:

• explicit
• conspicuous
• machine-readable

The Panton Principles come right out and say "explicit", and "machine-readable" is largely covered because the recommended licenses are available in machine-readable versions (though I'd have preferred to see that actual phrase in the text of the Principles). What's missing, to my mind, is "conspicuous". The point of Open licensing is to enable and promote re-use, so it's important to make your license as obvious as possible to potential users. This might seem trivial, but I think it bears spelling out.

My own Open Data mantra is:

• where are the data?
• can I have them?
• what can I do with them?

and again, the PPs are 2 for 3 by my count. The licensing covers what I can have and what I can do with it, but there's no mention of where I can find it in the first place. When we're talking about a database, the question doesn't arise since the license is in the same place as the data. But if we're talking about data which underlie a published paper, those data are very often not in the same place as the paper, even if the license is there. So it's important to make sure that your data are available: find or build them a stable online home and then let potential users know where it is. There's not much point in placing something in the Public Domain if the only copy is on your desktop. I'd have liked to see an explicit discussion of storage, access and signposting in the Principles... though come to think of it, this is really a different (and enormous) set of questions. So perhaps "conspicuous" covers this as well, and the missing Principle is simply that there should be a highly visible link to the license and the data themselves in every place where they are used, mentioned or otherwise likely to be encountered.

Of course, there are always unresolved questions no matter how carefully you craft your Declarations and Statements and Principles -- which is why the OKF has wisely built a companion tool, the Is It Open Data? web service. This is a brilliant way to remove ambiguity once and for all, on a case by case basis, by making public enquiry into the openness or otherwise of specific data sets. You can browse previous enquiries, so as to avoid redundant questioning of data owners; and naturally, recipients of multiple enquiries can use the service in a different way, simply linking to the record of their first response by way of answer to subsequent queries. Searchability might be a concern once the database of enquiries starts to grow, but that functionality can be added as needed. A central public service for asking questions about data availability and archiving the answers could go a long way towards improving access to data, simply by making clear the level of demand for Openness, and the degree to which supply falls short.

... sometimes someone puts together a programme that means you just have to shift the rest of the world around to make sure you can get there. Lisa Green and Hope Leman have put together the biggest concentration of speakers in the Open Science space that I think I have ever seen for the Science Commons Symposium -- Pacific Northwest to be held on the Microsoft Campus in Redmond on 20 February. If you are in the Seattle area and have an interest in the future of science, whether pro- or anti- the "open" movement, or just want to hear some great talks you should be there. If you can't be there then watch out for the video stream.

Along with [Cameron Neylon] you'lll get Jean-Claude Bradley, Antony Williams, Peter Murray-Rust, Heather Joseph, Stephen Friend, Peter Binfield, and John Wilbanks. Everything from policy to publication, software development to bench work, and from capturing the work of a single researcher to the challenges of placing several hundred million dollars' worth of drug discovery data into the public domain. All with a focus on how we make more science available and generate more innovation. Not to be missed, in person or online...

If you have any interest at all in Open Science (and why on Earth would you be reading me, if you didn't?), you should make every effort to attend this symposium. I'm a bit skeeved out by its being held on a Microsoft campus -- actually, I'm a lot skeeved out, and if it were any other lineup I probably wouldn't go for that reason alone. But this is simply too good to miss. Seriously, do yourself a favor and be there if you possibly can.