petermr's blog

The Wellcome Trust is one of the most valuable organisations in Open Access. It’s blazed a trail of clarity through the Open Access jungle. Very simply:

• WT require their grantees to publish Gold CC-BY Open Access
• They provide money to make sure it happens.
• They have a clear transparent record of they spend.

It’s a model that all funders should emulate.
So, yesterday, Robert Kiley announced a spreadsheet of Open Access spending. A heroic effort. Every paper is listed with the publisher, the title and the details. Here’s an excellent review: http://epriego.wordpress.com/2014/03/20/wellcome-trust-apcs-towards-an-open-access-serials-crisis/ . Cameron Neylon has tidied up the spreadsheet
Neylon, Cameron (2014): Wellcome Trust Article Processing Charges by Article 2012/13. figshare.
http://dx.doi.org/10.6084/m9.figshare.964812
and some of the OKFN have also contributed volunteer labour.
What follows is not contrived. I really did pick the first article in the spread sheet.
So I go to the first record (PMC ID, Publisher, Journal, Title, APC paid by Wellcome)
“PMC3378987,  Elsevier,  Academy of Nutrition and Dietetics, “Parent support and parent mediated behaviours are associated with children’s sugary beverage consumption”, £2,379.54
I put the title into Google (most people do this rather than using publisher searches) and get: http://www.andjrnl.org/article/PIIS2212267211019447/fulltext

 
 
Let’s zoom in:

How strange … I thought this was a list of Open Access papers. That you didn’t have to pay for.
Since I haven’t got the contract between WT and Elsevier I shan’t cast any aspersions. Maybe Wellcome got the details wrong? Maybe they are happy for Elsevier to charge for Open Access?
Or, just possibly, Elsevier have got some explaining to do.
UPDATE: Jan Velterop has tweeted that it’s available at http://www.sciencedirect.com/science/article/pii/S2212267211019447 as “Open Access funded by Wellcome Trust”. Of course it’s claimed by Elsevier as :”Copyright © 2012 Academy of Nutrition and Dietetics. Published by Elsevier Inc. All rights reserved.”.
So Robert Kiley thought he was paying for everyone to be able to read the article in the journal. He wasn’t. He was paying for the article to be put behind an Elsevier paywall.
So I invite the “Director of Access and Policy” at Elsevier to explain. And it’s NOT satisfactory to say “Oh you can get it for free at Science Direct”. Most people will look in the journal. And many haven’t a clue about ScienceDirect.
 
 

If you are confused by “Open Access” you are in good company. So am I. It’s horrendous. Here Danny Kingsley and the  Australian Open Access Support Group (AOASG) have created an excellent series of blog posts on Open Access. They are readable and authoritative. After it you will be less confused and more authoritative. But it will still be horrendous. My comments below.

You may be getting bored with my posts. I’m sorry but I am continuing to report injustice and that is never boring. Elsevier is still apparently behaving illegally and they don’t care.
Here’s a story from today. Thanks to Sheffield for sending this in.

 
 
The tweets point to the paper http://www.sciencedirect.com/science/article/pii/S1053810013001396 which was authored at the University of Sheffield.  I and the rest of the world interpret this as follows:

• @OpenAccessSheffield has first-hand or good second-hand evidence that the paper came from Sheffield and that the authors paid for Open Access with a CC-BY licence.
• Elsevier failed to provide it, so they complained
• Elsevier changed it to Open Access but failed to change the licence.

But apparently #elsevier think it’s OK to provide a different licence:

 
Yes, Director. This is the whole point. Sheffield asserts it is NOT CC-BY-NC-ND. They paid for CC-BY and you repeatedly failed to provide it. You are still failing to provide it and apparently still don’t care.
Even if Sheffield is wrong (and I doubt it), #Elsevier have still mislabelled the article as (C) Elsevier, All rights reserved
 

 
 
Do you have any idea how demeaning it is to pay for something and repeatedly be fobbed off with excuses and substandard products. And have to try to explain to the funders that actually it is the publishers’ fault. A publisher who either doesn’t understand the difference between CC-BY and CC-NC-ND or doesn’t care.
I would encourage Sheffield to (a) take this to their local trading office and (b) write to David Willetts as I have done.
My suspicion is that a large number of “open access” articles are mislabelled. Only the authors (and possibly Elsevier) know which they are. So please keep reporting them to show that Elsevier’s behaviour is now intolerable.
 

 
 
Here’s the university of Cambridge’s reply to my request about barriers to  content-mining imposed by publishers. It’s exactly what I had asked for. (It’s a pity that several other Universities decided it was too much effort.) See earlier blog posts for background and the actual question here: I’ll annotate briefly, comment on each issue and summarise
 
 
 
1. [How many restrictive contracts]. All license agreements between publishers and libraries set out permitted uses of the data being
provided and restrictions on use. Most do not specifically mention text mining but include a phrase such
as ‘Users may not … abridge, modify, translate or create any derivative work based on the licensed
product(s)’. Of the publishers with which the University Library has signed licenses, 11 are known to
place restrictions on text mining or have licenses that could be interpreted as having restrictions.
PMR: very clear, especially final sentence. I suspect this is true of most universities.
2. [when was first contract with restrictions? ] The University Library has been signing license agreements in large numbers with publishers since
the mid-1990s; it would require substantial research to answer this question with any certainty in
relation to any specific restriction such as that relating to text or data mining. Accordingly the
information you have requested in this question is refused under section 12(1) of the Act. The
University has estimated that the time required to locate, retrieve and extract the information would
considerably surpass 18 hours of staff time charged at £25 per hour, and therefore that answering this
question exceeds the appropriate limit of £450 as set out in the Freedom of Information and Data
Protection (Appropriate Limit and Fees) Regulations 2004.
PMR: probably not a useful question to ask
3. [How often has University challenged a restrictive contract?]. Never, and hence no information is held to answer this question.
4. N/A.
5. [Has university raised this in a committee]. Minute 6 of the Minutes of the Journals Co-ordination Scheme Steering Committee meeting on 21
June 2012 states: “Mr Morgan raised the question of text/data mining rights and requested that this be
made explicit within the NESli2 standard model licence when deals are re-negotiated. Mrs Jarvis
commented that this matter had been raised previously via ERWIG with JISC Collections and will be
under discussion at further meetings.”
PMR: Excellent. Peter Morgan has done an excellent job of driving innovation in Cambridge.
6. [How many researchers have approached university to do content mining?] Four approaches (two individual researchers; one pair of researchers working on a single project;
one research group) have been made to the University Library to contact a publisher to request
data/text mining. Two of the four requests were rejected by the publishers.
PMR: Really valuable data. Publishers refuse requests. The STM association has said publishers are extremely willing to cooperate. This proves otherwise. Yes, it’s a small sample but enough.
7. In the two cases under question 6 above where the publisher did not object, one did not impose
conditions and advised on technical procedures to gain access to the dataset for text mining purposes.
The other agreed on condition that the University Library should update its license agreement and that
the researcher should register with the publisher for text mining and agree to its text mining conditions
via a ‘click-through’ agreement.
PMR: very valuable insight. This allows a followup up question as to who were the publishers, etc.
8. [Have researchers go unfavourable responses from publishers.] Never, and hence no information is held to answer this question.
9. The University Library has never advised a researcher that they should desist from mining but has
made them aware of publisher policies. Subscription access to a resource has never been removed
due to a breach of license terms relating to text mining.
10. [Policy on click-through licences].  No such policy exists, and hence no information is held to answer this question.
11. [Policy post Hargreaves legislation.] No such policy exists, and hence no information is held to answer this question.
12. [Will university refuse to sign restrictive contracts after Hargreaves?] The University Library, on behalf of the University, will continue to license electronic information
resources to support research and teaching, negotiating the best possible terms of use with publishers
and pressing to establish the right to carry out text and data mining within license agreements. It will do
this both as an individual organisation and in collaboration with consortia negotiators.
 
PMR: Overall this is very helpful and it’s a pity that others didn’t take this view. I have transmitted some of this to my MP and David Willetts. It’s through information of this sort we can support them to change policies.
 

Nearly 20 working days ago I FOI’ed 10 Russell group universities about their contracts and practices of content-mining and their relations with researchers and publishers. Over half have responded – the others have a day or two left. In summary the responses vary from very helpful to useless.
FOI is a formal process. Universities have to reply but they don’t have to do my work for me. In an ideal situation they would have all the information to hand and simply post a link to the relevant documents or summarise in a sentence.  There’s an allocation of up to  450 GBP / 18 hours for the effort required.It’s acknowledged that some answers may require a lot of work and the universities can refuse to do this unless they are paid extra. It’s potentially reasonable to either ask me for extra money above the 450 GBP or to bundle unanalysed documents and ask me to extract my own information. It’s also allowable to say that the university doesn’t hold the information (but they have to be convincing) or to refuse because of confidentiality. I was expecting a wide range and I got it. As I have only done 40% of Russell I will refine my questions (e.g. I should probably have routed questions to Edinburgh’s central University library).
Let me again thank MySociety for WhatDoTheyKnow. It’s just brilliant. I don’t have to manage my email – all the correspondence is trivially available and shared with the whole world. I can see, for example, that Durham have got 6 hours to reply in. (Some universities routinely leave it till the last day). Here’s the summary – the links are to WDTK. In general the “Successful” are what can be reasonably given under FOI. There is a bit of SirHumprheyism in some – especially in arguing that it would be too costly to ask everyone in the whole university. But others have been positively helpful. My summary:
Cambridge: First class honours (not because I’m in Cambridge)
Cardiff, Exeter: Upper seconds
Bristol, Edinburgh, Kings: Played the “money” card, whereas they could have been more helpful with little effort. Technically OK. but…
Glasgow. “We don’t know anything”. Oh dear. I offered them a resit. They still failed.
Here are the links. You can also add comments for the universities:

I have sent the following letter to my MP:

Dear Julian
LEGISLATION on EXCEPTIONS TO COPYRIGHT and PUBLISHER CLICK-THROUGH
LICENCES
Please copy this letter to David Willetts.
I know you have been championing the Hargreaves reforms and note David
Willetts’ statement [1] and comments from JISC [2]. I am concerned by
the  JISC report last week [3] that the legislative process was delayed
again. I hope you can reassure me that this is days, not months.
I am very concerned by recent activity from major scholarly publishers
the effect of which may be to nullify much of the legislative benefits.
Elsevier have recently provided a “click-through” licence  for
researchers which gives “permission” to carry out content mining if
their Universities have signed a general agreement. I have done a
personal analysis [4,5,6] of this “licence” (with which my commenters,
including Charles Oppenheim, agree) and find it to be even more
restrictive than the current practice. In effect Elsevier is the sole
arbiter of what is allowable (e.g. licencees are forbidden to do
anything which competes with Elsevier’s data products). There are
indications that other publishers will do similarly, perhaps through a
generic site such as Crossref. Since a major part of Hargreaves is to
create a new wealth creating data industry this will be stifling.
I am also concerned as to whether University libraries will fail to
adopt the Hargreaves opportunity and instead continue to accept
publisher licences. I have written to several Russell group
Universities under FOI. The complete replies are not yet in but none
have formally prepared for Hargreaves. All currently sign restrictive
contracts and, for example, the University of Cambridge has had half
its requests for content-mining refused by publishers.
Assuming you can push through the legislation my suggestion would be to
empower HEFCE, through JISC, to contact all UK universities and make
clear that they should not sign restrictive contracts, not allow
click-through contracts, and prepare the ground for widespread content
mining through Hargreaves. Since some publishers might challenge what
will be legal activity I hope you can support researchers (like me) to
assert our legal rights.
For note: I have been funded by the Shuttleworth Foundation to develop
an open platform and community for mining scientific papers in detail
and I would be very interested to discuss your personal requirements
(e.g. we could have a daily alert of every DNA quadruplex with
molecular mass, ions, sequences, etc.)
With best wishes,
Peter
[1] http://www.ipo.gov.uk/hargreaves-copyright
[2]
http://www.jisclegal.ac.uk/ManageContent/ViewDetail/ID/2077/Summary-of-the-Potential-Impact-for-HE-and-FE-of-the-Hargreaves-Review-16-January-2014.aspx
[3]
http://www.jisclegal.ac.uk/ManageContent/ViewDetail/ID/3479/Further-Delay-to-Changes-to-UK-Copyright-Law.aspx
[4]
/pmr/2014/02/06/content-mining-elseviers-tdm-why-researchers-and-libraries-should-think-very-carefully-and-then-not-sign-1/
[5]
/pmr/2014/02/06/elseviers-tdm-terms-tac-can-they-force-us-to-copyright-data-2/
[6]
/pmr/2014/02/07/contentmining-and-elseviers-terms-the-small-print-absolutely-prevents-responsible-science/

Lest you think this blog is solely concerned with bashing Elsevier, it tries to be impartial. Most of the time I get my targets from the puffing of publishers on Twiiter flaunting their “open access offerings” and it just happens that the worst offender by far is Elsevier. But there are others with completely unacceptable activities. Here I am very grateful to Thomas Munro, who writes (/pmr/2014/03/17/elsevier-admits-it-has-been-mis-selling-open-access-and-will-be-contacting-mis-sold-customers/#comment-152848 ):

Strangely, I’ve just found that BMJ do the same thing, e.g. this “CC BY-NC” article
http://www.bmj.com/content/348/bmj.g1741 (under “Request Permission”)
links to a Rightslink page that quotes $36.80 for an individual to make 10 photocopies. Furthermore, no-one may “reproduce in full or in part any material that promotes the use of baby milks / infant formula without reflecting current knowledge and responsible opinion and be restricted to scientific fact and not imply that or create a belief that bottle feeding is equivalent to breast feeding” or use a reproduction to promote sunbeds (!)
Shall we call it CC Attribution-NonCommercial-NonXerox-NonInfantFormula-NonSunBed?
 

Oh dear! Where to begin. Well here’s the screenshot:

 
BMJ has immediately and grotesquely broken the terms of CC. However much we may wish to outlaw sun beds  or smoking, you CANNOT use CC-BY licences to do this. CC does not discriminate against fields of endeavour.  The software community wrestled with this over thirty years ago and agreed that licences could not discriminate. If you wish to make poison gas using my chemical software I cannot use licences to stop you.  So, BMJ , you will need to find another way. (I use cursing).
And I am delighted to have readers’ contributions – keep them coming.

Elsevier’s Director of Access and Policy (was “Universal Access”) has commented on my blog posts.

Dear Peter:
We appreciate your concerns about this matter, and I can assure you that Elsevier is taking the issues you have raised seriously and that we are working as diligently as possible to have these resolved. Some elements we were able to address in short order last week, including disabling the links from open access content on ScienceDirect to Rightslink.  Please also note that we rapidly committed to reimburse anyone who has made an unnecessary payment. We have completed the investigation for open access content published by us under a CC-BY license and there appear to be 11 customers who may have been impacted and to whom we will reach out in the coming days.  We need a little more time to look into other license types.
We’re currently working on a more complete and official update that will reaffirm our commitment to this shared journey to implement the UK government’s OA policy, while noting that there will be some bumps on the road. But rest assured we take smoothing these out very seriously.
Thank you,
Alicia
Dr Alicia Wise
Director of Access & Policy
a.wise@elsevier.com
@wisealic

This confirms that Elsevier has been mis-selling Open Access products. The DoAP does not give sufficient detail to say exactly what so I don’t regard this reply as satisfactory. Given also that part of the problem appears to be a seriously broken IT infrastructure it cannot be accurate to use it to find problems. It is unclear whether the 11 customers were authors with mispublished articles or purchasers of rights through Rightslink.    This problem is of similar seriousness to faulty electrical goods where responsible companies will advertise in the national press. I would therefore have expected Elsevier to make a public announcement of the following sort:

“Elsevier informs its authors that some of its “open access” products have been unfit for purpose and will issue a full or partial refund in the cases where:

• an article has been hidden behind a paywall
• an article has not been labelled as “open access”
• an article has been labelled “(C) Elsevier” or similar
• an article has been linked to “Rights and permissions” which imply that there is limited or no permission for re-use

We ask any author who believes they have been disadvantaged to contact the Director of Access and Policy.
Elsevier also informs purchasers of rights from Rightslink that they may have paid for rights which were already granted. Anyone who believes they have made such payments is asked to contact the director.”

From my own discussions I would expect the scale of the problem to be higher than 11. The only way I could be convinced would be an independent audit.
 

Robert Kiley has just published a spreadsheet of every single paper they have sponsored as paid (“Gold”) open access. http://figshare.com/articles/Wellcome_Trust_APC_spend_2012_13_data_file/963054 . This is brilliant. Every funder should do this.
So off I went to see if the papers were really Open Access. Have the publishers fulfilled their bargain? I chose an Elsevier one – at random – #471 in Robert’s list. It’s CC-BY (hidden at the bottom of the 50-page paper) but still “All rights reserved”.
Can I get free and Open re-use. Off to Rightslink to see if I can get PERMISSION for translation (I’ll have to do the translation – this is just a permission tax). Elsevier ID S0010945213002037

 
32 USD for each reprint… So for a CC-BY article Elsevier are still claiming copyright (breach of copyright law) and mis-selling.

I am so excited to be supported by Shuttleworth – I’m still running round in circles working out where to go. The reason is that the support is massive. Not just money, but mentoring, experience, etc. And especially the heartbeats.
Projects benefit massively from having a “heartbeat”. This is something that happens at regular intervals, rain or shine. It keeps focus. It helps redefine deadlines, etc. It should not be judgmental – you say what you have done, intend to do. If you consistently run late then that’s an objective observation. It’s then up to project management to take action. If you over-deliver again management should decide/advise what to do.
I have set myself ambitious tasks in my Fellowship. Some of this assumes that I will get synergy from others I discover. I’m used to long projects – Chemical Markup Language (CML) has been 20 years. But I am also used to projects which take off in weeks (cf SAX where we – communally built an important Internet standard in a month). So I’m always alert to miracles of human collaboration.
But realistically often TTT (Things Take Time, Piet Hein). I don’t know what the speed of the ContentMine will be. I’m getting together a small formal team which I hope to announce soon and I’m making strong contacts with several organisations (more later).
The fantastic thing is that the Shuttleworth Foundation puts massive resources into support for projects. This is beyond anything I have been used to. For research councils there is often no formal support during the project – “here’s the money, tell us in 4 years’ time what you have achieved”. There’s usually little contact with other grantees – often they are your “competitors”. With JISC it was much more intimate – and I appreciated it. JISC had “programmes” with “programme managers”. They have regular contact and I appreciated this. “How can we help with your project? Would you have synergy with organisation X?” and they would run communal days which might be theme or programme-related. For example Dave Flanders got us together in B’ham to discuss “Open Scholarship”.
But Shuttleworth goes beyond this. There are at least three heartbeats:

• regular two-weekly one-to-one Skypes with staff (Helen Turvey and Karien Bezuidenhout). So last week I had an initial meeting with Helen in Southwark and cam away with such as wonderful feeling of support.
• A regular Fellowup every week. Here present and past Fellows Skype in confidence – always at the same time. I missed the first (I was in the air) but just managed part of the second. (As I was chairing a meeting on Open Government Data at the time I had to find a chunk in the tea interval. I’ve already received an offer of help from one fellow.
• a twice yearly physical get-together. I’m really looking forward to this.

I’ve also got a mentor Fellow (“buddy”) whom I already know.
I shall come across problems and opportunities and I now know that I have got a support structure where I can ask the Fellowship and Trust – in confidence – “Help!” Because they’ve probably come across the problems several times already – getting partners, raising money, building a team, legal challenges, political, etc.
My initial actions are roughly:

• build a team
• seek partners
• plan workshops
• write software and infrastructure
• advocacy and publicity (mainly blogging).
• political (as we are at a very active tie with the Hargreaves legislation)

That’s a lot to keep in order . My heartbeats will be critical.