PhiloWeb 2012 keynote: Stevan Harnad by PhiloWeb

UPDATE. I got feedback suggesting that part of principle 2 was inappropriate at this stage and I agree. So I have struck through parts in this post. There is merit in changing emphasis at such an early stage in the process. This document is subject to revision – that’s part of the point of open discussion.

We – in the OKFN – have been spending some time on Etherpads and skype putting the principles of Open Content Mining. Yesterday we met on skype and decided that we’d done sufficient to take this to the world and get feedback and enhancement. Naomi Lillie (OKFN) will post the full version later and Peter Suber will link to it. This blogpost is an introduction and I’ll quote the central points.

Kinder Scout (from Wikipedia) Fuaigh Mòr (Wikipedia)

Let’s start with another historic area of rights – the right to roam. This is a 20^th C movement in many countries to assert that everyone has access to land, whether or not it is privately owned. It’s a good analogy. The fundamental ownership of land id critical, political and often poorly defined. In the 18/19^th Century Scotland suffered the http://en.wikipedia.org/wiki/Highland_Clearances – where the residents of the land were thrown out – killed, emigrated, – the lands “improved” with sheep and the lands now “belong” to landlords. But there is a traditional right of access to these lands regardless of actual “ownership”. Wikipedia (http://en.wikipedia.org/wiki/Freedom_to_roam ) says:

The freedom to roam, or everyman’s right is the general public’s right to access certain public or privately owned land for recreation and exercise. The right is sometimes called the right of public access to the wilderness or the right to roam.

Not everyone shares the same view as to what these rights are or even whether they exist. I have been thrown off Scottish land by a gamekeeper with a shotgun, even where there was a legal right. But just because not everyone agrees on the rights doesn’t mean they don’t exist.

So we believe that there is a right to mine the scientific literature and we have expressed this as:

The right to read is the right to mine.

That’s our assertion of the fundamental rights. In the 20^th Century the people asserted their right to roam. We are asserting the people’s right to mine. This is a simple political statement – like “everyone has a right to a fair trial”. Because the publishers[*] – like the 19^th C landowners dispute this right we have to fight for it. The UK has had a series of fights for rights including freedom of speech, trial by jury, freedom from slavery, etc. Sometimes people went to jail, sometimes they died for these.

But we must fight. An extremely relevant example is the mass trespass at Kinder Scout (http://en.wikipedia.org/wiki/Mass_trespass_of_Kinder_Scout) , WP:

The mass trespass of Kinder Scout was a notable act of willful trespass by ramblers. It was undertaken at Kinder Scout, in the Peak District of Derbyshire, England, on 24 April 1932, to highlight that walkers in England and Wales were denied access to areas of open country. Political and conservation activist Benny Rothman was one of the principal leaders.

The trespass proceeded via William Clough to the plateau of Kinder Scout, where there were violent scuffles with gamekeepers. The ramblers were able to reach their destination and meet with another group. On the return, five ramblers were arrested, with another detained earlier. Trespass was not, and still is not, a criminal offence in any part of Britain, but some would receive jail sentences of two to six months for offences relating to violence against the keepers.

The mass trespass marked the beginning of a media campaign by The Ramblers Association, culminating in the Countryside and Rights of Way Act 2000, which legislates rights to walk on mapped access land. The introduction of this Act was a key promise in the manifesto which brought New Labour to power in 1997.

So it’s a long struggle. Am I suggesting a Mass Trespass of publishers? That may depend on readers. But the same tensions are there as 80 years ago – an unjust control of access and the need to change the system by breaking the law. And we have a long tradition of noble lawbreaking – often it is the only way that we change minds and therefore laws. There is usually a debate as to whether change should come by legal means or – in today’s language – “occupying” and “pirate” action.

So I repeat:

The right to read is the right to mine.

This isn’t a negotiated position. It’s not a summary of current practice. It’s a statement of a fundamental right that we must fight for.

Yesterday we agreed that we could not at this stage list the “how” of Open Content Mining (OCM). That comes later. It will probably be filled with subjunctive clauses – this is a difficult and complex area. The right to roam has to yield to national security and rare species. It may or may not have to yield to personal privacy – a difficult area. So the right to mine will have to take account of the current law and decide what can be done within it or what needs changing (e.g. Hargreaves). It may require a definition of “fact”. It may requires cases. It could take some time. But that does not mean we cannot NOW assert the right.

So here’s the core of the principles. We’d welcome others being involved. But I repeat, this is not a negotiation – it’s drafting something we expect to stand for decades or longer. Much of it needs commentary and redrafting – particularly IMO section 2. We don’t want to rush these principles, but we do wish to kickstart the process.

Principles

Principle 1: Right of Legitimate Accessors to Mine

We  assert that there is no legal, ethical or moral reason to refuse to  allow legitimate accessors of research content (OA or otherwise) to use  machines to analyse the published output of the research community.   Researchers expect to access and process the full content of the research literature with their computer programs and should be able to use their machines as they use their eyes.

• The right to read is the right to mine.
  

Principle 2: Lightweight Processing Terms and Conditions    

Mining  by legitimate subscribers should not be prohibited by contractual or  other legal barriers.  Publishers should add clarifying language in  subscription agreements that content is available for information mining  by download or by remote access.  Where access is through  researcher-provided tools, no further cost should be required. The right  to crawl is not the right to use a publisher’s API for free, however,  when access is through publisher-supplied programmatic interfaces, the  fees should be transparent and per-api-call.  Processing by subscribers  should be conducted within community norms of responsible behaviour in the electronic age.

• Users and providers should encourage machine processing.
  

Immediate feedback suggested deleting part of this section and I agree.

Principle 3: Use

Researchers can and will publish facts and excerpts which they discover by reading and processing documents.  They expect to disseminate aggregate statistical results as facts and context text as fair use excerpts, openly and with no restrictions other than attribution.  Publisher  efforts to claim rights in the results of mining further retard the  advancement of science by making those results less available to the  research community; Such claims should be prohibited.

• Facts don’t belong to anyone.
  

[I meant to blog this earlier but I have been spending time on writing content-mining software rather than the continued depressing struggle with reactionary commercial publishers of #scholpub.]

On 2012-05-24 the Guardian (a/the mainstream liberal daily newspaper in Uk published an article in its main news pages on content-mining from scientific #scholpub articles. In the paper “It’s a useful research tool so why forbid it?” (p14), online http://www.guardian.co.uk/science/2012/may/23/text-mining-research-tool-forbidden

Text mining: what do publishers have against this hi-tech research tool?

Researchers push for end to publishers’ default ban on computer scanning of tens of thousands of papers to find links between genes and diseases

Byline: Alok Jha, Science Correspondent

Alok was the person who promoted “Academic Spring” on the front page of the Guardian last month. He contacted me and others, especially Robert Kiley from Wellcome Trust. Robert and his Wellcome colleagues have made a massive contribution to free scientific information – without Wellcome we would have much poorer involvement. And as sponsors of UKPMC Robert is at the frontline of content-mining – he knows firsthand how hard it is to get any help from the publishing industry[*].

The coverage included stories from Casey Bergman + Max Haeussler, Heather Piwowar and myself – detailing carefully and accurately our major ongoing difficulties. Some snippets:

All of them [above] needed access to tens of thousands of research papers at once, so they could use computers to look for unseen patterns and associations across the millions of words in the articles. This technique, called text mining, is a vital 21st-century research method. It uses powerful computers to find links between drugs and side effects, or genes and diseases, that are hidden within the vast scientific literature. These are discoveries that a person scouring through papers one by one may never notice.

It is a technique with big potential. A report published by McKinsey Global Institute last year said that “big data” technologies such as text and data mining had the potential to create €250bn (£200bn) of annual value to Europe’s economy, if researchers were allowed to make full use of it.

Unfortunately, in most cases, text mining is forbidden. Bergman, Murray-Rust, Piwowar and countless other academics are prevented from using the most modern research techniques because the big publishing companies such as Macmillan, Wiley and Elsevier, which control the distribution of most of the world’s academic literature, by default do not allow text mining of the content that sits behind their expensive paywalls.

Absolutely correct.

Any such project requires special dispensation from – and time-consuming individual negotiations with – the scores of publishers that may be involved.

“That’s the key fact which is halting progress in this field,” said Robert Kiley, head of digital services at the Wellcome Trust. “For a lot of people, though there is promise there, the activation effort is just too great.”

Exactly. My research has been set back 2-3 years by fruitless “discussions” with publishers.

Asking for permission from publishers is an option, though time-consuming. The University of British Columbia (UBC) researcher, Heather Piwowar, was trying to map the ways scientists use and share papers.

She was eventually contacted by Alicia Wise, Elsevier’s director of universal access, who convened a conference call with Piwowar, a UBC librarian and five Elsevier colleagues. That conversation led to permission for UBC researchers to text mine the Elsevier journals to which they already had access.

Piwowar said: “It takes a lot of time and a lot of energy and doesn’t scale at all. To me it’s a good result because now I have access to things I didn’t have access to before and also it will also hopefully drive change by people saying, ‘This is not an OK way to build on our scholarly literature.’”

The colossal waste of time is clear. Elsevier want me to negotiate with them and the Cambridge University Library. I have to tell Elsevier what research I want to do. The library has better things to do with its time. So do I.

And it’s technically completely unnecessary. I can access the articles I want by standard means. It’s a pinprick in the daily Elsevier downloads. It’s sheer FUD to suggest I will crash their servers. I don’t want ZIP files from them through a special API. I already have what I want. All I need is Elsevier to say they won’t sue me.

Wise said that, in principle, her company was happy to enable text mining for its content. “We want to help researchers deepen their insight and understanding, we want to help them to advance science and healthcare and we want to be able to do that in ways that help realise the maximum benefit from the content we publish. Text mining is clearly a part of this landscape and it will continue to be and we’re keen to support it.”

“In principle” means nothing. In the comments AW described

Elsevier is leading the research information industry to enable text mining.

NO! BMC and PLoS are leading it. I can mine them – as much as I like and I can’t mine Elsevier at all.

We provide text mining solutions to an array of customers, and we also enable researchers to text mine our content for themselves. This is all done through licensing, which is highly efficient and easily scalable.

So efficient and scalable that I have got nowhere in ca. 3 years. So efficient that we need 5 Elsevier staff for one researcher.

We began partnering with the University of Southern California in 2007 to enable researchers in its Neuroscience Research Institute to content mine and we now have agreements with about 20 universities around the world.

Wow! 20/1500 universities in 5 years. Just over 1%.

We also serve researchers in a broad array of commercial organisations. Earlier this year we announced our acquisition of Ariadne Genomics and QUOSA, companies that both provide state-of-the-art text mining services to improve researcher productivity. We continue to invest to develop an array of text mining ourselves, and we offer other tools through collaboration with partners such as the UK’s National Centre for Text Mining. We are also working with other publishers to ensure that text mining is possible regardless of who has published it or where it is located.

These are all words. I am still not allowed to text-mine. And it is Elsevier who makes the rules – in most science it’s God who makes the rules, but here it’s Mammon. I will write a blog on Elsevier and Helpfulness. “Elsevier is a helpful publisher” is similar to a British bank which advertises “helpful banking”. Think of “helpful banking” whenever you think of Elsevier.

Back to the positive.

So what Alok has done is massive! To get national coverage at this level is a huge boost to the legitimacy of our effort. It means the issue is now clear to everyone and cannot be ignored as a minor fringe activity. The UCSF declaration for Open Access (still not mandatory and therefore of very limited practical effect) mentioned mining. Funders are starting to promote mining. UKPMC is fully aware of its huge potential – the dam is only maintained by publisher lawyers and publisher lobbyists in Capitol Hill (US).

So I have been aggressively tooling up for when I am allowed to mine the scientific content. The Guardian article acted as a trial in the court of public opinion and I think the publishers have very little support there.

But I am starting with BMC. Who knows, maybe there is enough hidden science in just 5% of the scholarly literature?

Today we continue developing our Manifesto on Content Mining

P.

[*]Yes, I exempt PLoS, BMC, and lots of worthy society publishers

This post notes some reflections from a recent meeting of the Community of Practice of the  Ontario Ministry of Agriculture, Food and Rural Affairs (OMAFRA) / University of Guelph’s Knowledge Translation and Transfer (KTT) group. My role in this group is that of open access consultant. The OMAFRA / KTT group is doing some very interesting work in the area of developing intellectual property practices to support innovation, including both open access and patenting. Researchers include academics and also grower groups.

One of the projects involves growing the Ontario vegetable crop research repository in the University of Guelph’s ATRIUM repository. The University of Guelph’s Ontario Agricultural College produces a lot of agricultural research, much of which has world level impact, particularly in the areas of corn, fusarium, and pesticides. Much of this knowledge is currently available only in unpublished research reports stuck in filing cabinets. If these reports were digitized and made available through ATRIUM, the research would be useful to many people – including Ministry staff for developing policy and local farmers and gardeners wondering whether to use black plastic on their strawberries.

One of the challenges to developing the repository is dealing with rights issues. Much of this research is owned by growers’ groups, who conducted the research for their own community. As agricultural entrepreneurs, the growers will want to retain an edge for competition and so are likely to want to retain commercial rights. Similarly, faculty members at Guelph own their own IP and may want to retain the rights for commercialization when applicable. Strategies to address these issues could include such tactics as defensive publishing.

My thoughts so far as shared in the meeting:

Engaging the growers’ groups in open access is a strategy that I would highly recommend in this situation. Don’t just ask for the license to their works, rather do some workshops or provide information to link people to some of the many open access resources that are already available to them. A message of people everywhere are sharing their work; won’t you join us? strikes me as a message that is a little bit easier to listen to than won’t you share your work? Include open access peer-reviewed journals on agriculture, of course – but don’t neglect to mention some of the high-quality magazines written largely by people similar to the growers’ groups, such as BC Grasslands. Focus on agriculture for sure, but not necessarily just agriculture – farmers and their families are people too, and are as likely as anyone to benefit from all the freely available health information or enjoy the many free texts, movies, and music available from the Internet Archive. Flickr can be a good resource for developing marketing materials, and open government resources can be useful, too.

One challenge for farmers in this area is that many still rely on dial-up access. This suggests to me another avenue for illustrating the benefits of open approaches. It can be difficult for people in rural communities to get the rest of us to pay attention to their issues (such as lack of broadband) and hence to gain political support. This is one area where the internet creates the possibility for a more level playing field; a rural newspaper can create an online presence with the same potential audience as an urban newspaper, and rural individuals, families and community groups can similarly create an online presence with the same potential impact as urban people.

Some potential venues for information sharing include practioners’ peer-reviewed journals using tools such as Open Journal Systems – although the growers’ groups might be more interested in using social networking tools like ning. 

Our Community of Practice is just getting started! Watch for further posts on this topic.

I’m spending an exciting two days in Berlin helping the OKFN/P2PU prepare their School Of Data (SoD) course/s. I’m sure this will turn out to be a seminal event in both Internet education and advancement in “data wrangling”. Here’s the initial announcement – http://blog.okfn.org/2012/02/08/announcing-the-school-of-data/ . “The School will be a joint venture between the Open Knowledge Foundation and Peer 2 Peer University (P2PU). ”

There’s a huge need for skilled and inventive data wrangling. This is a mixture of technical knowledge and knowhow and the “course” will cover both. We are working out the granularity of the “course” – almost certainly a collection of smaller units, generally self-paced but with some clear timelines. P2PU has had considerable experience in this – for example partnering with Mozilla on web skills.

Here’s Laura Newman – the course coordinator – getting our thoughts organized and photographed, and here’s Rufus Pollock and Stiivi Urbanek hard at work planning the details.

Stiivi has put together a great “architecture” for the technical side of the course which goes from acquiring data, to cleaning, filtering, repurposing and presentation. We have a strong sense of pipeline, where course participants take a problem from start to finish, using the appropriate skills are each stage. We are presenting this round “challenges” – we take a theme which everyone can relate to and go all the way from finding the data to drawing conclusions.

The course structure and participation is flexible and controlled – there is no hierarchical distinction between teachers and leaners – we are all a bit of both. We expect information to flow from and to the course.

The overall components (stages) – which have largely crystallized in our planning – are

• Data sources
• Discovery and acquisition
• Extraction
• Cleansing, transformation, and integration
• Analytical modelling
• Data mining
• Presentation, Analysis, publishing and packaging

And an overarching subject of “data governance”

To analyse a particular subject a participant needs to go through the processes above, although not all will be needed for a given problem/challenge. We call this process a “journey”, where we visit the different stages on a planned itinerary. Many courses will be organized like this – and the first we have designed is “What is unique about my country?”

In this participants (perhaps working in teams) will find and extract information about their country, clean, fliter and integrate it and finally present answers to this very general question (which requires comparison with other countries).

In an orthogonal fashion, participants will also study a particular stage in depth. In the journey metaphor, this is like spending your time in one place, finding the different ways of tackling it. So one early topic will be “Crawling and scraping” – there are several different tools, approaches and problems.

There’s a real buzz! Over 300 people have signed up and we had an IRC meeting yesterday with 30 – who are very keen to be involved and contribute. Lots of great skills and ideas.

Much more later – on a regular basis – as this is an important part of my life.

I recently asked “What’s the difference between Elsevier and British Gas?” I didn’t get many answers (it would be nice to have a greater response so I could highlight ideas other than mine). The question could also have replaced “British Gas” by “Virgin Trains”, “Scottish Power”, “East Anglian Water” or even “Lloyds Bank”.

The answer is that the others are all, to a greater or lesser extent bound by regulation. They have a legal duty to:

• Ensure the quality of service
• Limit prices

Scholarly publishing is in a bizarre and completely unhealthy marker where there is no effective market regulation of price, there is no quality control (the quality of #scholpub is awful compared to other e-products on the web and hasn’t changed in 20 years. ) We have NO IDEA what the true costs of publishing a paper are, or what they could be if the market operated.

Acta Crystallographica E publishes the highest quality papers in science. It’s a data-only journal and doesn’t completely scale to other journals. It charges 150 GBP for Gold Open Access and makes a margin. They have built their own authoring system which every crystallographer uses and the papers are full of checked, semantic data and there is high-quality peer review. It’s difficult to extrapolate but I think a figure of 500 GBP would be the MAXIMUM cost of an efficient scholarly publisher. I’d like to see the high price publishers challenge this.

Yesterday I was asked by a journalist (I won’t spoil their story) to comment on the UK Finch report. This hasn’t formally reported but there are some open readable minutes at http://www.researchinfonet.org/wp-content/uploads/2012/01/Minutes-mtg-2012-04-272.docx and I was asked to comment on what I thought of the pricing , market, etc. I don’t have a strong view on Finch, but it says:

The Working Group first considered the tables in the annex, which were founded on modelling undertaken for the Heading for the Open Road report. It was noted that the ‘central case’ was a starting point under which APCs were set at a cost-neutral level for the HE sector in the UK of c£1,450 per article, with an assumed take-up rate of 23.3% for OA publications. All the tables therefore use that as a starting point, and vary the costs according to a series of different assumptions – some of which are obviously more realistic than others. The variability is determined by four factors: (i) the level of APCs, (ii) the level of take-up of the gold option, (iii) the difference between levels of take up in UK and rest of the world, and (iv) the proportion of APCs to be met by authors outside and within the UK for jointly-authored papers. The Group observed also that the £18.7m saving from subscription charges does not take account of ‘stickiness’ in a transitional shift from subscription to APCs – which is liable to take a significant amount of time. Such a transition implies additional costs.

I haven’t read the annexe and I cannot see how they can actually assess the costs since almost no publishers analyse and publish them. Some publishers have argued that costs can approach 20,000 USD because of high rejection rates. This is a typical example of an unregulated market. It’s like saying “we don’t have enough capacity on our buses so we are going to throw most passengers off and charge the others a huge amount to make our profits”. It’s a sign of a broken market.

A typical example of how inefficient the industry is and how unresponsive to costs is that most publishers send the manuscripts off to be retyped – this is an appalling admission of lack of reaction to the 21^st century. It’s like having to send Amazon a snail mail to order something. It’s because Amazon broke the model that we have efficient, price-competitive market of goods. If the academic sector wished to reduce costs of Gold OA they should create a system with author-side cost reduction. If I was given the option of paying 1450 GBP for APC or 500 GBP if I created it in NLM DTD XML I’d go for the latter. The NLM (which publishes Pubmed) is a world authority on publishing and far more efficient than publishers. It has been highly innovative and the only brake on progress has been the relentless destructive legalisation against it and restrictive practices imposed by major toll-access publishers. That’s why we cannot get access to content-based search, not because they can’t do it.

Anyway I wrote the following for the journalist. It echoes what I have written here:

“What I am concerned about [and what I intend to blog about as soon as I have time] is the lack of regulation in this market.  In almost all transactions, whether author->publisher or publisher->reader there is no price-sensitive market. There is little market pressure on publishers to bring down costs, nor to produce better products. (Scholarly publishing is one of the very few sectors to be completely unaffected by the web – the product is an electronic copy of what was done 20 years ago). There is even less market force in the hybrid Gold model where publishers can charge what they like with no regulation – it is simply up to the funders or authors to pay what is demanded. Moreover the products offered are often not significantly different from Green – there are no rights of re-use and in some cases not even of copying.

In areas such as transport, energy, banks, public services and many others the government regulates the market. Providers have to work within negotiated margins and provide an agreed level of service. None of this pressure is put on publishers. The market often resembles personal vanity products where only the brand matters and cost of production is irrelevant.

My view is that any Green/Gold model will be a seriously suboptimal model until all the current cost (10 billion USD/yr) can be brought funder/author-side. This desperately needs regulation and strong leadership from bodies – probably governments and major funders. I don’t think Finch has addressed this at all – you cannot be convincing unless you demand a change of control and do the budgeting properly.

I believe that even at 1500 GBP per paper this represents a seriously overpriced market. I think it might be brought down by bringing in public contractors / purchasers as is done in Brazil, I believe. Nothing could be more inefficient than leaving market forces to libraries in 10,000 scattered uncoordinated universities.
”

So I am not getting excited about Finch unless the government (Willetts) does. AFAICS Finch says “we want a mixed Green/Gold model with the emphasis on Gold. We aren’t putting money in. We aren’t imposing regulation. We are not controlling prices related to costs.” And of costs it’s only one country.

#scholpub is now, at its worst , a vanity market such as fragrance or mineral water. The price is vastly higher than the cost. You ask what you can get, not what it costs. There is large, wasteful marketing, there is large and wasteful investment in technology and lawyers to prevent access.

So what’s the difference between Elsevier and Chanel? Not much. They are both unregulated.

Oh, and stop thinking of publishers as collaborating partners. Alicia Wise on the GOAL Open Access mailing list asks “what can publishers do to help”. She asserts publicly that I don’t trust her. Actually I trust her completely. I trust her to behave like a middle manager public relations officer in “Customer Relations” for British Gas, or Scotrail or whomever. She is there to maximize profits for the company. And part of that is preserving the current pseudo-monopolies. I trust he to continue to try to defend that. And offering help is a well-used strategy.

And she can trust me to challenge almost everything that Elsevier does, says, and more importantly doesn’t do.

Stevan Harnad is dismayed that Elsevier has introduced a catch-22 int their Green regulations. It’s convoluted (well-designed Catch-22s are) and says something like “you can deposit Green, but if your institution mandates it then you cannot”. Stevan feels this is a breach of trust and that Elsevier should change it. I say that until this is regulated by a body with teeth we shall continue to have these games played by the publishers. If I travel to somewhere via London on British trains the price is higher. The cost is not higher.

Think of Elsevier, Nature, Wiley, Springer, etc as gas, transport, telecoms, etc. They have no more reason the be loved or hated than those.

The sick part is that the trains have to pay for their fuel (and a lot else). In #scholpub we GIVE the publishing industry the content.

I has been always been possible to provide a link to the OpenDOAR entry for a given repository using its persistent numeric identifier. However, these link URLs have hitherto been rather long and cumbersome. Now we have implemented simpler and shorter links based on REST principles. For example:

http://opendoar.org/id/227/ – OpenDOAR entry for ‘Nottingham eTheses’ (ID: 227).

OpenDOAR identifiers and the corresponding new link URLs can be found in the relevant entries in the database.

We think that this new feature is likely to be of particular interest to users of the OpenDOAR Application Programmers’ Interface – http://opendoar.org/tools/api.html.

Peter

The Hungarian government has ended (without apology) its FUD campaign against philosophers that had been critical of the government (the “Heller Gang”). Investigation over, no evidence of wrong-doing, all charges dropped, case closed. Yet the FUD campaign succeeded in doing its intended damage. That’s what FUD is for…

And the Hungarian Academy of Sciences has suitably disgraced itself, for betraying its historic mandate and refusing to stand up for its members — as announced by it President, Joseph Palinkas (former minister in the present Hungarian government).