Version classiqueVersion mobile
OpenEdition Books

White Paper — Open Science in a Digital Republic

Scientific and Technical Information Department - CNRS


Minutes from the hearings

Texte intégral

UPMC, Jean Chambaz and Paul Indelicato, 9 June 2016


For the UPMC:

  • Jean Chambaz, President

  • Paul Indelicato, Vice-President for Research and Innovation

For the Cabinet Alain Bensoussan:

  • Alain Bensoussan, Barrister, specialised in law relating to advanced technologies

  • Laurence Tellier-Loniewski, Barrister, Director of the Intellectual Property Unit

For the DIST CNRS:

  • Renaud Fabre, Director

  • Charlotte Autard, manager in charge of the ISTEX Investments for the Future project


1The CNRS, in conjunction with the Cabinet Alain Bensoussan, decided to draft a White Paper on “Digital Science and Law” in response to issues raised by the legal framework covering the ISTEX Investments for the Future project ANR-10-IDEX-0004-02 ( This project raises many legal questions due to its potential in terms of TDM, interdisciplinarity, content aggregation and its aim of making its databases explorable.

2This White Paper aims to propose a legal framework for scientific data in order to address the concerns of the scientific communities (data publication and uses, laws governing science platforms, text mining, etc.), and thereby contribute to the French Digital Republic Bill.

3Initially, the Cabinet Alain Bensoussan and the DIST will establish working groups and conduct hearings to gather information about the community’s uses and the current state of exploitation practices. 

4This information about uses and exploitation practices will enable the CNRS and the Cabinet Alain Bensoussan to establish a matrix concerning the relevance of analyses and practices which will be set against the current normative framework in order to assess gaps in the law and make suggestions for the Bill.

Copyright and the law as applied to science

5Currently, there is no legal text governing science, apart from copyright law. Science has been bound by copyright as regards its legal framework, its evaluation and its dissemination. Copyright covers not only past science (articles) but also future science (analyses of results).

6Copyright was designed to privatise the form of a piece of creative work by attaching it to its author, while enabling publishers to make money and contribute to culture through its dissemination. Later, things were reversed and now authors are somewhat under the thumbs of publishers, which collect 90% of the value of their creative work.

7Today, copyright appears unsuited to digital science. The aim of the White Paper on Digital Science and Law is to propose solutions for the creation of a new law governing science.

8It is extremely difficult to separate considerations concerning copyright for scientific data from the general issue of copyright, which has been the subject of intense debate for a long time. It may be more productive to focus on the notion of data and data rights, which is a distinct issue from copyright and can therefore be addressed without any misunderstanding.

9Scientific data are constantly evolving, and although data science will never replace the scientific method, data and even more so the reuse of data are at the heart of this new approach.

10Certain fields of research are based on the collection of large amounts of data on a specific subject or gathered from human activities (public, governmental, health data, etc.). These data have value for the research project in question but can also be reused for other research projects. It must be established who actually owns the data collected: the person who collected the data and labelled them, the organisation, or the laboratory that stores them?

11The data must be considered a public asset which can be reused freely. If we could manage to define a specific status for scientific data (or data for scientific purposes), then copyright would cease to be useful.

The specific status of data

12Jean Chambaz and Paul Indelicato differentiate between two types of data:

  • observable data;

  • constructed data: the collection of several pieces of data to which specific methods or algorithms are applied to allow observations to be made from a specific collection of data.

13The intellectual process that makes sense of the data collected remains the intellectual property of the researcher and the funding organisation, but there is no reason the data produced during the research process should not become a public asset with a specific status allowing for its reuse. Creating a specific status for data could stimulate research activity for the benefit of society as a whole.

14The way data is collected and used depends on the discipline concerned. For certain disciplines, collecting and sharing large amounts of data are crucial for consolidating the discipline, whereas in other areas, such as experimental science, reusing the data collected is not always relevant.

15Paul Indelicato points out that, depending on the discipline, for scientific data to be of use it must be accompanied by the method and procedure that produced the data.

National (or European) Data Library

16Today, the majority of data collected during research projects is escheated or deposited in dispersed databases or directories. If data are to be deposited in a specific National (or European) Library, then the associated publications must be deposited with the data.

17A European Library would have greater value for disciplines such as medical research, for which sharing data on a national level is not particularly relevant.

18It could take a long time to set up a European Data Library; Jean Chambaz therefore recommends that this factor be included in the current debate concerning the Juncker Plan, and endorses the CNRS’s conclusions on this point.

New legal aspects to be taken into account in the scientific process

19Today, moral codes, religion and ethics have varying degrees of impact on new rights in the scientific field.

20If we want to argue in favour of the creation of a global warehouse for free data, we must provide justification that the warehouse is a common universal asset. It is essential that scientists should be able to carry out TDM on scientific data. If this possibility depends on publishers, researchers will be subject to strict controls with no apparent limits.

21Publishers are currently worried about the dangers that abusive TDM could represent concerning large community databases (data extraction, pirating, etc.). However, the solution to this problem involves educating users, not controlling access. Establishments, organisations and laboratories must educate and train their staff, researchers and students concerning the use of TDM as a scientific, economic and cultural tool. To enable TDM, the data must be labelled intelligently so as to be reusable and widely shared.

22Paul Indelicato stresses that the conditions for reuse must be attached to all data subject to mandatory deposition.

Economic exploitation of the data

23Raw data has no specific value in itself; collecting the data is not a neutral act and is carried out within a specific framework, whereas the information itself does not contain all the ideas for how it can be exploited. It could be argued that the person who collected the data should benefit from all the potential exploitation of the latter. This fact would incite researchers to deposit data in a library and would give great added value to the establishments/organisations that finance the research. This would give value to the collection, processing and labelling of data in order to make them easier to exploit.

24The notion of a scientific and/or economic embargo could be established for researchers who wanted to impose a delay before the data were deposited in the library (constructed data).

25Jean Chambaz points out that the free reuse of data by the scientific community may achieve consensus, but the law must also provide for the reuse of data by companies and others who benefit from the use of public research. A balance needs to be found concerning the reuse of data for economic and scientific purposes.

26For example: the collection of data linked to the behaviour of SNCF (rail) travellers could be used by a scientific team for a sociological study as well as by the SNCF to optimise new services.

27Companies could have access to the data to develop their activities on the condition that a return for society as a whole is legally provided for. A clause concerning access to the data by private companies could provide for a fee on the exploitation of the data so that the National (or European) Data Library could finance itself, e.g. via a tax on data use.

Legal instrument for data

28In terms of social acceptability, a data item may be freely available for reuse when this means reuse by the scientific community. As regards reuse by companies, a distinction should be made between observable data (free to reuse) and constructed data. As regards constructed data, research teams could use a legal instrument (patent, licence, etc.) to maximise ROI (return on investment) on the data they have collected and labelled.

29In a limited number of cases, this legal instrument could justify an embargo period of five years (to help write off the costs). This period must be tested to ensure the system provides enough incentive to motivate researchers to deposit their data.

Ethical data

30Processing and labelling data creates a new mission for public researchers. They must not only conduct scientific research but also deposit the data processed, which represents an additional cost in time and human resources.

31Researchers must therefore not only conduct research but also ensure the digital transfer of data for the benefit of the community.

32Alain Bensoussan notes that a new concept is currently coming to the fore; the notion of “ethical” data. This means data that has been “fairly” collected, data that allows for sustainable development.

33Data for sustainable development should not be destroyed but tracked and deposited in a database.

34Scientific communities should be educated and trained in adopting a sustainable development approach as regards the data they collect for their research projects.

35Researchers must understand that processing and depositing their data does not represent an additional cost but new added value. The protocol concerning the processing and depositing of data must be promoted as a valuable part of the researchers’ work. This must become an integral part of their job, just like the assessment and publication of their research.

The CNRS Scientific Board: Bruno Chaudret, Claire Lemercier, François Bonnarel, François Tronche, 30 June 2015


For the Scientific Board’s working group:

  • François Bonnarel, CNRS Engineer, Strasbourg Astronomical Data Centre (CDS)

  • Bruno Chaudret, President of the Scientific Board of the CNRS

  • Claire Lemercier, Senior Researcher in History at the CNRS, Centre for the Sociology of Organisations (CSO), Paris

  • François Tronche, CNRS Research Director, Paris-Seine Institute of Biology, CNRS, INSERM, UPMC

For the Cabinet Alain Bensoussan:

  • Alain Bensoussan, Barrister, specialised in law concerning advanced technologies

  • Laurence Tellier-Loniewski, Barrister, Director of the Intellectual Property Unit

For the DIST:

  • Renaud Fabre, Director

  • Charlotte Autard, manager in charge of the ISTEX Investments for the Future project

Overview of the approach

36In the context of the ISTEX Investments for the Future project, the CNRS Legal Affairs Department (DAJ) commissioned a legal consulting firm to provide project support, at the request of the DIST. The Cabinet Alain Bensoussan was selected to support the Executive Committee in establishing a secure legal framework for the project. This collaboration has already enabled the studies conducted by ISTEX on the Digital Republic Bill to be taken into consideration.

37This collaboration has led the Cabinet Alain Bensoussan to support the DIST of the CNRS in drafting a White Paper aimed at building a set of references, which have been largely reproduced in the Digital Republic Bill on open access to scientific data and the principle of text and data mining (TDM).

38The French Government’s Digital Strategy draws on the CNRS’s conclusions aimed at higher education and research in terms of digital technology for science, and reasserts open access and TDM as key elements in this approach.

39The Digital Republic Bill will be presented before the month of July. As such, the White Paper is of great importance. The conclusions drawn from the information collected about uses and practices from interviews with the Scientific Board and other partners (CNN, UPMC, COEPIA, University of Strasbourg, ISTEX Executive Committee) will be included in the parliamentary debate concerning the Bill. The DAJ and the DIST, supported by the Cabinet Alain Bensoussan, will work together to present the positions of the research community supported by the Scientific Board of the CNRS.

40The initial feedback from the Scientific Board must be sent to the DIST (for transmission to the Cabinet Alain Bensoussan) before the end of August 2015 to ensure enough time to draft the White Paper before the filing of the Bill for the parliamentary session at the beginning of October.

41The Scientific Board’s contribution will help focus the debate on the White Paper during the next round of hearings. Minor modifications may be made to the text proposed by the Scientific Board following its next plenary meeting on 24 or 25 September.

Presentation of the objectives

42The goal of the collaboration with the ISTEX Executive Committee, guided by the DIST, the DAJ of the CNRS and the Cabinet Alain Bensoussan, is to find a legal framework for the ISTEX project. A preliminary analysis was made during two CNRS Ethics Committee (COMETS) meetings.

43Discussions concerning ISTEX led to the creation of two projects:

  • a “limited-scope” project concerning ISTEX and the law governing platforms and TDM (more or less free use of scientific data);

  • a “wide-scope” project concerning a more general legal framework of scientific and technical information (STI): a law governing science. Both projects are based on a scientific approach for the use of data.

44The Scientific Board’s working group is now invited to give its opinion concerning these two “limited-scope” and “wide-scope” projects.

45As regards the “wide-scope” project, the DIST and the Cabinet Alain Bensoussan have already obtained the opinions of two leading experts: Laurent Cytermann, the Master of Requests (Maître de Requêtes) for the French Council of State (Conseil d’État), and Alain Abecassis, Head of the Department for Strategic Coordination and Regions.

46The White Paper focuses on the opinions of the members of the Strategic Board in their capacity as scientists in order to understand the legal requirements linked to their fields of research.

Publications and data

Definition of data

47Data can be sub-divided into five categories:

  • Raw: available in their original state before any human processing;

  • Instrumentalised: obtained using a given instrument (e.g. a telescope in astronomy). The person who controls the instrument controls the data;

  • Analytically interpreted: resulting from calculations or processing. Ownership must be established between either the person who supplies the data or the person who supplies the algorithm;

  • Scientific data: interpreted by the human brain (E=MC2). Establishing links between data. This often concerns data that have to be interpreted to support a theory;

  • Data about data, or metadata: all the information about and relationships between pieces of data that make it possible to interpret the data itself.

48These lie at the heart of several usages:

  • Open Data (open governance of data);

  • Open Access (possibility of free access to data (or not) with embargo periods for certain fields of research);

  • Open Process; 

  • Open Format;

  • Open Use;

  • Open Business.

49Finding the right definition for the data can sometimes be complicated, for example, in the case of a photo of a work of art.

50At first glance, the photo is a piece of instrumentalised data but the object captured by the photo exists outside the photo itself and has its own rights, i.e. those of raw data. There can be legal conflicts regarding the reuse of raw data represented in picture form.

51Numerous scientific fields work with data that are not produced by science, i.e. raw data; the uses of these raw data must therefore be precisely determined.

Differentiation between data and articles

52An article that is written within the framework of state-subsidised research studies is financed using public funds. Based on the work conducted, researchers communicate using an international standard: the scientific article. Researchers thus become the owners of the articles in their capacity as authors, protected by copyright law.

53Researchers pay to have their articles published and then pay again to have access to the journals in which their articles and those of their colleagues are published. Most often, they do not have the possibility of performing TDM.

54This observation leads to the following questions: Is the existing system still suited to the opening up and sharing of science? Are private articles still the most appropriate way of disseminating science? Is copyright still adapted to the needs of science?

Compendium of uses

Practices and sharing: Principle of Fair Use (TDM)

55In biology, digital publishing has been generalised and article searches are performed via platforms developed by academic institutions, such as the National Institutes of Health (NIH, USA), providing free access to article summaries. A fee must generally be paid to obtain access to the complete articles, with the transfer of copyright to the publisher being a common practice. It should be noted that the NIH has objected to this practice and proposes open access, via PubMed, to an unformatted version of any article published by a publisher describing work funded by the NIH. Open access is growing rapidly.

56The production of research articles published in digital format is not only accompanied by the production of raw digital data but also of equipment (collections of DNA, anti-bodies, cells, apparatus, software, etc.) and living organisms (microorganisms, strains of mice – selected or genetically modified) used in the scientific process presented in the research article. Depending on the type of production, their ownership is subject to copyright or the practices of the supervisory authority. Let us take the example of a genetically modified mouse: the ownership depends on the practices of the supervisory authority and any potential revenue is divided between the inventors, the authority and the laboratory. When a published article describes research that produces raw digital data or biological material, the publisher generally asks the author to sign an undertaking to donate the data or the material to any researcher from a public organisation who might request it. This practice is essential for validating results by reproduction and is facilitated by the existence of international platforms providing free access to data and distribution of the biological material.

57TDM on texts is not a priority in most fields of biology as a discovery tool, although it is as regards documentary collection. TDM on data is becoming increasingly commonplace. Free access to these data is widespread, as happened concerning the human genome.

58Above and beyond the issue of mass data, several publishers, including Nature, are considering implementing a system allowing access to the raw data used to create the graphs in an article via their websites. While this will enable readers to verify that results have been interpreted correctly, it raises the question of who owns the data and how ownership is transferred. This suggestion that publishers could extend their control to include data is worrying and may represent a new obstacle to accessing science.

59When we compare uses, astronomy would appear to be the best organised research field in terms of data and access interoperability.

60The community reached a consensus for establishing an embargo period for disseminating and reusing instrumentalised data. This embargo period comes into force only on publication of the article, and applies after its publication according to a variable time period prior to free reuse.

61François Bonnarel points out that astronomy is a research sector in which the commercial value of data is often zero, and this therefore facilitates data sharing. Moreover, the need to exchange data is driven by the wide variety of observation techniques (linked to different wavelengths) potentially available for a given object.

62François Tronche points out that in biology data access is free for the academic community, sometimes with a one- or two-year embargo period; this period does not depend on the publication date since the embargo begins when the data are discovered.

63As it stands today, there is no specific document that stipulates the embargo period. In astronomy, the “ownership” period is defined based on the instrument or the complete project (several instruments).

64The embargo period gives the researcher an exclusive right to his/her results and data for publication purposes.

65As such, we are faced with a two-sided economy: a donation-based economy (data available freely to researchers in the same community) and a market-based economy (access to publications).

66Claire Lemercier notes that researchers must sometimes work with large groups such as Amazon to have access to large storage areas (example provided by Pierre Binetruy during a previous meeting), but always in exchange for something (payment, access to the data, etc.). Most of these large groups also have research laboratories.

67If we plan to make the data available to the scientific community, how can we verify that they are not used by these private research laboratories to make a profit? How can we establish whether the reuse of the data is for scientific purposes, for profit-making or for other purposes (educational, for example)?

68If a historian decides to use works of art for research or training purposes, these uses must be differentiated from the use of the data to publish “fine editions” of works of art (to make a profit).

69In order to define the possible fields of data reuse, we could create a principle of Fair Use. There is great demand for the reuse of data by the scientific community and in education (the right to quote and reproduce for research and training purposes).

70The concept of a right to practice TDM could therefore be extended into a principle of Fair Use.


71Exchange is a key part of research. If this is restricted by controlling access to data using logins (Application Programming Interfaces, or APIs) or payments (subscriptions, article processing charges, etc.), scientific progress may be slowed down.

72Fair Use provides for the right to use or quote a text on the proviso that the text is an extract, printed between quotation marks, with the author’s name indicated. However, Fair Use is not valid for pictures. Pictures cannot be reproduced even with quotation marks, the source and the name of the author. This is also true for extracts of sound recordings.

73This practice was relevant for analogue technologies but is not suited to the digital age.

74Copyright remains sacrosanct in France. If we want to move things forward, we cannot seek to create an exception to copyright. For the moment, the focus must be on the data item and not on the article. In this way, science could one day return to an exception to copyright through the concept of “data right”.

75It is therefore crucial that a consensus be found between the different scientific communities in order to foster free data exploitation. If the data item is not free, it must be covered by the principle of Fair Use to allow for its reuse.

76Claire Lemercier notes that the principle of Fair Use as regards sound extracts or pictures would be possible only for data produced by scientists. The problem in HSS is that this can involve the reuse of data that are not “scientific” in nature (for example, it is possible to study objects found during excavations, communications on company websites, or pop songs). Other data used in HSS cannot be made freely accessible because they are subject to private data protection or non-disclosure clauses. In certain cases, as in the field of public statistics, a solution was found by making the data anonymous.

77Today, some publishers ask HSS authors to attach their data to their articles, particularly in economics; however, most HSS data cannot be published (personal data, psychological test data, photos of archive documents that archivists allow researchers to take only on the proviso they do not disseminate them, etc.).

78The HSS community would favour the creation of exchange platforms but the legal issues need to be addressed for this to be possible.

79For example, we could create a platform that allows free exchanges between researchers by making the data anonymous, as is the case for statistic exchange platforms (projects are under development for more “qualitative” data, but this is far from simple).

80The platforms (datasets about a given project made available on a given computer) could be recognised as “essential infrastructures”.

Legal amendments expected

Community expectations

81The Cabinet Alain Bensoussan, commissioned by the DIST, would like to interview the Scientific Board about these expectations with a view to drafting a text on science law and a text on platform law.

82Today, digital law concerning science is not clearly defined. Science produces platforms, but there is not yet a platform law. Establishing a science law is a major challenge as regards the recognition of scientific community practices. Today, digital science requires a specific law that allows it to evolve in terms of its uses and practices without being restricted by copyright or abusive publishing practices.

83While the universal principle of science is well established, the advent of digital technology means we can no longer continue to have a donation-based economy that coexists with a market-based economy.

84With the CNRS itself, there is a paradox between the open access policy and the policy of the “technology transfer” service.

85Claire Lemercier points out that science must address the issue of misappropriation of results and data. Could this appropriation become legitimate? When is data appropriation considered abusive – not only appropriation by publishers but also by other external players?

86The White Paper’s approach must be based on the rules provided by the scientific community.

87Claire Lemercier points out that, while this approach is valuable, particularly as regards preventing misappropriations and increasing access possibilities, there are two clear risks:

  • Standardisation: what degree of standardisation would a science law involve? The different practices of different scientific communities must be respected since these practices are generally there for good reasons.

  • How non-scientific outcomes might be taken into account by digital science law: how could large groups such as Amazon and Google (and other companies in other fields, such as insurers, who also have their own research centres) be prevented from freely appropriating scientific content? There are certain types of data that scientists dream of being able to share freely but which would pose serious ethical problems if exploited for profit-making purposes. A data item is not intrinsically scientific or non-scientific. It is the use that is scientific, but how can we establish what constitutes scientific use?

88The law governing science could provide for:

  • a principle of Fair Use;

  • the free exchange of all elements in a scientific process between scientists;

  • an embargo period before data publication/dissemination. Researchers would thus have priority access to the data for a given period (depending on article publication rates or exploitation initiatives).

Construction of digital science law

89Digital science law must be built; we can therefore propose content but this must be supported by scientific communities in order to be valid. If we create a legislative text that is supported by the scientific community as a whole, its adoption will be quicker.

90François Tronche points out that everything that is described in an article that is not patented can be reproduced by anyone. Many researchers do not file for patents so that their findings can be disseminated to a broad public.

91Moreover, it would be possible to envisage not actually publishing, and thus not letting the article or data item become private property, while not restricting access either.

92Publishers are starting to become interested in metadata. Metadata must also be protected against misappropriation by publishers.

93If services to categorise and use metadata in a smart way were private or fee-paying, François Bonnarel believes researchers unable to access these tools would be less competitive than those who could afford to privatise their metadata.

Embargo period

94The ownership period cannot be determined in numbers of years because the duration of projects can vary from a few months to a few decades. We could consider an embargo period as being up to the date of publication of the article. This period would be linked to the notion of the project’s end, as long as it is possible to determine the project period (start and end). Alternatively, the embargo period could be defined by an agreement.

Creation of an alternative publication model

Journals financed by public institutions

95If we cannot change copyright, there are alternative models that could be developed to enable publishing without the problems of misappropriation.

96In France, there are HSS journals (often financed by public institutions and based on an open access (OA) model), for example, those present on the OpenEdition platform, which are managed by universities and the CNRS. These new OA modes of publication, which are not subject to misappropriation, could become models for the future.

Impact factor

97In France, the problem with OA journals is the impact factor. In many disciplines, the assessment of researchers in France and Europe is primarily based on their publications in journals with a high impact factor. This system is actively promoted for the assessment of research. Journals that attract articles are therefore those with a high impact factor. Since the impact factor depends on the journal’s selectivity, the system can have perverse effects and certain OA journals have been able to increase the number of fee-paying articles published in the year during which their impact factor was high. This increased their profits, even if it lowered their impact factor the following years.

98This type of assessment based on impact factor is changing, even in the United States. Scientific texts by French researchers are formatted with this assessment method in mind.

Information sharing

99Although, for the moment, in most fields of biology, researchers see the point of being able to consult articles, but not of being able to “mine” them, the community actually needs free access to articles as quickly as possible. In biology, sharing also involves sharing manufactured material. Many publishers want the community’s access to the material described in the article to be guaranteed so that anyone can reproduce the experiment. Furthermore, several publishers are currently implementing a system for accessing raw experimental data summarised in article tables or graphs.

100Although the biology community appears to share the most, this sharing does not solely concern data but also the material produced within the framework of the article.

101In HSS, the community needs to protect itself from data appropriation and to have free access to articles, thus reducing the cost of access to information. TDM is also important; this concerns access to a large body of resources. Data sharing is not yet sufficiently widespread; there is a vacuum in terms of culture and human resources as regards sharing. If a law governing public sciences and platforms is created, it will perhaps enable HSS communities to implement more widespread data-sharing platforms based on the model of statistical data platforms.

102In HSS, it would also be useful to establish a principle for open data and sharing that would continue after the end of the project; however, for most data, this should probably be limited to sharing among scientists, with all the problems of definition that this entails.

103In astronomy, the community shares a lot of data, but the reuse rate remains low. Algorithms must be linked to articles published in OA so as to make it possible to check the result and render it reusable.

104The working group concludes that the common concern is to enable the free sharing of information between scientific communities. To achieve this, an intelligent means of sharing data must be found.

Software sharing

105In astronomy, as regards scientific results, when we talk about OA we refer to the sharing f instrumentalised (or raw) data on which the algorithm (that makes the data usable) is built. Could the algorithm and the software also be shared?

106François Bonnarel says that the sharing of software is not only desirable but already occurs. However, there is not yet a general, standardised system for exchanging software. Nonetheless, there is a whole range of interoperable software packages in the field of virtual observatories, although this represents only a small proportion of the software programmes used in astronomy.

107François Tronche points out that in biology, not just the associated material but also the software can be shared, if it is described in the article. Most of the time, the software developed by the academic community is freeware. Certain large centres or academic organisations develop their own software in open access mode.

Specific forms of sharing practices and knowledge

108Many platforms are created and then disappear according to needs.

109EMBO (an international research and publishing organisation) was the first to implement a service providing access to the raw or processed data used for graphs.
(Thomas Lemberger –​)

110François Bonnarel notes that a recent paper, listed on the INSU website (National Institute for Earth Sciences and Astronomy) talks about articles that are no longer based on observations made within the framework of research studies but on archived data. In terms of the reuse of forgotten or unexploited data, this trend is growing in astronomy, where researchers sometimes do not make new observations but publish articles using open access archived data.

111François Tronche adds that, in biology, a large proportion of a laboratory’s experiments are often subcontracted, within the framework of partnerships or service contracts. In practice, the person who acts as subcontractor can do this for no fee if they are cited as co-author of the article, or for a fee otherwise.

University of Strasbourg, Paul-Antoine Hervieux and Françoise Curtit, 10 July 2015


For the University of Strasbourg:

  • Paul-Antoine Hervieux, Deputy Vice-President for Partnerships with EPSTs and local authorities

  • Françoise Curtit, CNRS, Responsible for the “Open Access” mission at the University of Strasbourg

For the Cabinet Alain Bensoussan:

  • Alain Bensoussan, Barrister, specialised in law relating to advanced technologies

  • Laurence Tellier-Loniewski, Barrister, Director of the Intellectual Property Unit

For the DIST:

  • Renaud Fabre, Director

  • Charlotte Autard, manager in charge of the ISTEX Investments for the Future project

Overview of the approach

112The CNRS, in conjunction with the Cabinet Alain Bensoussan, decided to draft a White Paper on Digital Science and Law in response to the issues raised by the legal framework covering the ISTEX Investments for the Future project ANR-10-IDEX-0004-02 ( This project raises many legal questions due to its potential in terms of TDM, interdisciplinarity, content aggregation and its aim of making its databases explorable.

113The new law on science has a specific exception concerning copyright: the legal status of scientific results and notably data and metadata.

114This White Paper aims to propose a legal framework for scientific data in order to address the concerns of the scientific communities (data publication and uses, science platform law, text mining, etc.) and thereby contribute to the French Digital Republic Bill. Science is preparing the terrain for the digital law, by working on tangible points upstream.

115Initially, the Cabinet Alain Bensoussan and the DIST will establish working groups and conduct hearings to gather information about the community’s uses and the current state of exploitation practices. The close association with universities and other organisations will enable hearings to be conducted with key witnesses and ensure pluralistic expression regarding the changes being prepared as part of the Digital Republic Bill.

116This information about uses and exploitation practices will enable the CNRS and the Cabinet Alain Bensoussan to establish a matrix concerning the relevance of analyses and practices which will be set against the current normative framework in order to assess gaps in the law and make proposals for the Bill.

117Contributions to the White Paper will be anonymous and the people interviewed will have total control over whether their names are quoted or not.

118The DIST and the Cabinet Alain Bensoussan will re-contact the people interviewed and will submit an initial text to the University Presidents for approval.

119Today, the Cabinet Alain Bensoussan (commissioned by the DIST) would like to record the opinion of the University of Strasbourg concerning potential draft proposals and recommendations.


120Paul-Antoine Hervieux points out that in academic communities there is an almost total lack of awareness of copyright and its implications, regardless of the type of document or data. The issues raised by the digital law represent a huge undertaking in terms of training.

121The White Paper aims to collect the opinions of scientific communities, which are aware of any deficiencies. The DIST and the Cabinet Alain Bensoussan would like to know if the same rights concerning scientific results and data could apply equally to different scientific communities.

122According to Paul-Antoine Hervieux, a law concerning data and results is becoming essential in the light of developments in the digital sector. Uses are changing and the paradigm for research data (BSN10) is undergoing a transformation. Researchers and academics are beginning to realise the value of their data. Once we start talking about value, we need to start thinking about rights. We are now in a world where private businesses predominate and are increasingly interested in data with a view to commercialising their uses.

Legitimate appropriation and misappropriation

123The University of Strasbourg is acknowledged for its research in the field of chemistry. The university owns many resources, such as experiment data and databases in which a considerable number of properties concerning chemical reactions are recorded. The latter can be of great interest to private companies; for example, the pharmaceutical industry is increasingly interested in the catalogues of chemical reactions, with a view to reusing this information for its own purposes.

124These data must be protected. Scientific communities are not yet aware of this danger, notably in universities.

125Renaud Fabre highlights that we must differentiate between legitimate appropriation and misappropriation.

126A specific text concerning Open Science could define what constitutes misappropriation.

Business model for data

127Paul-Antoine Hervieux highlights that the University of Strasbourg has data in its laboratories that could be very valuable for the private sector. A business model could be created using the data and could be used to support and finance fundamental research. The data should be open, even to the private sector; however, to do this, a business model would have to be defined, specific rules established (a guide to data), or an easy-to-use tool developed for researchers.

128Research can be differentiated by:

  • financing through public funds;

  • financing through private funds.

129These aspects must be used to define a business model for making the data available to private organisations.

130One possible idea (as an outcome of the White Paper) would be to propose an A-to-Z, frequently asked questions aimed at researchers so they could find their way through the labyrinth of data law.

Embargo period

131Astronomical data are produced by specific instruments: instrumentation data. The community puts an embargo on their data for a one- or two-year period. Could the idea of an embargo period be shared with the University of Strasbourg?

132Paul-Antoine Hervieux answers that the key factor in science is the immediacy of research. Each data item has a certain life cycle, but the fresher the better. For science to advance, the embargo period should not be an obstacle to the dissemination of results. However, depending on the scientific community concerned and for various reasons (e.g. competition, assessment process under way, etc.), an embargo period may be introduced.

Software and algorithms

133To maximise potential data reuse, one must have access to equipment that can process it. When data are open and made available to communities, should the software and algorithms also be made available online?

134Paul-Antoine Hervieux says that researchers have a philosophy of competition today. Each researcher must publish a certain number of articles a year to remain competitive. With this in mind, researchers will not be particularly willing to share the results of their work, or the calculation codes that are at the heart of their research. It is possible to share information about the structure of the code (but not about the algorithm itself, at least not for a certain period of time) in an article describing the methodology used or in a journal that publishes calculation codes. However, the majority of researchers are wary of immediately publishing their codes and algorithms.

135Today, there are already journals such as Computational Physics that allow physicists to deposit their calculation codes.

136It would also be possible to deposit the executable programme in open access mode with a copyright on the code.

137If communities want to have an embargo period for the depositing of software and algorithms, a period of one to two years would be suitable.

Scientific publications and repositories

138Paul-Antoine Hervieux points out that the relationship with publishers is central to the concerns of researchers, particularly as regards TDM. TDM will be the ultimate research tool in the years to come.

139Publishers seek to protect themselves by controlling the possibility of TDM via Application Programming Interfaces (APIs), under contractual clauses (Creative Commons licences, restricting the number of “explorable” words).

140It is obvious that the principle of open access must comply with copyright, but a distinction must first be made between the exploitation of corpora and the reuse of data already compiled. We must differentiate between raw data and constructed data.

141When an institution has an open archive, it has the opportunity to implement its own policy. At a time when university establishments are becoming increasingly independent and must establish their own strategies, open archives and TDM are a major issue.

142The University of Strasbourg aims to build its own archive to store its scientific production and make it possible to locate it in a single place. This archive will give the university community in Strasbourg the right to consult and search through its own production with no lock-outs (unless there is an official lock-out, such as Sherpa, or an embargo).

143According to Paul-Antoine Hervieux, the CNRS should ideally take the position that all scientific production (that of the CNRS and its partners) should be deposited in a common archive (not necessarily a single one) with similar rules.

144As the contact person for the EPSTs at the University of Strasbourg, Paul-Antoine Hervieux points out that university–CNRS partnerships work well despite them having distinct policies. It therefore seems obvious that the same data could be shared with the same rights.

Possible actions

145An article in the draft text proposed by the Cabinet Alain Bensoussan could state that: “Data from scientific research and the associated results can be freely used and made freely available to scientific communities involved in the publications, with no restrictions, except for software, which may be made available to the community after an embargo period of one to two years.”

146Renaud Fabre points out that the purpose of the HAL platform is to incite CNRS researchers and organisations to deposit their publications.

147Paul-Antoine Hervieux explains that the University of Strasbourg’s major project to build an open archive of the university’s knowledge also involves developing a connection between the university’s archive and HAL so that publications are automatically deposited on the platform. The CNRS is supporting this project. The University of Strasbourg wants to focus on creating a repository of all the work produced by its researchers, essentially for promotion purposes, but also aims to comply with the CNRS and European policies via the automatic transfer of information to HAL.

Digital Republic Bill

148Françoise Curtit raises the issue of the calendar for the digital law in France and that of European Directives that are currently being amended. Will European law not take precedence over the French digital law?

149Alain Bensoussan points out that the President announced this regulation as creating a French law on digital issues in a similar vein to that of human rights. The Prime Minister supports this initiative and a text is already being studied by the Cabinet of the Secretary of State, Axelle Lemaire.

150The Research Code deals with the organisation of science but not with that of scientific data. The Digital Law could include an article that would amend the code on scientific research and take into account the White Paper’s recommendations.

151Renaud Fabre says that the Government’s digital roadmap was made public by the Prime Minister on 18 June 2015 and that the provisions in terms of science promote the development of open access.

152These provisions also cover the authorisation for text and data mining (TDM) and an amendment to the Intellectual Property Code is planned with this in mind.

153In both cases, the roadmap clearly sets out the requisite elements for open access and free browsing through scientific literature.

Open Process

154The issue of open process and of whether it should be possible to search freely through data highlights the question of legitimate and misappropriation. The DIST and the Cabinet Alain Bensoussan aim to analyse this issue with the representatives of the communities interviewed. The hearings have already improved and clarified the definition of this practice. The Scientific Board of the CNRS will give its opinion on these subjects in September.

155It is also essential to gather the opinions of major universities (such as Strasbourg and UPMC – University Paris 6) to have an overall view of the French research ecosystem.

Platform regulation

156The question is as follows: should platforms be regulated and qualified as “essential infrastructures” for scientific research?

157Paul-Antoine Hervieux believes it is crucial to establish a legal handbook for the use of platforms. This code of conduct would allow platform users to understand the regulations and uses and adapt their behaviour accordingly. The University of Strasbourg is already looking at drafting a document about platform use.

158Renaud Fabre states that “essential infrastructures” are defined by two conditions: they must be essential and they can be implemented only through the resources of public authorities (e.g. airports).

159Digital platforms for science are essential and can be a means only for sharing knowledge.

160For example, the TGIR HumaNum (Very Large Research Infrastructure for the use of digital resources in the HSS) currently has no specific status but fulfils several needs of scientific beneficiaries. Labelling it as an “essential infrastructure” would establish it as an upstream structure of public science.

161Paul-Antoine Hervieux agrees with Renaud Fabre about the need to recognise digital platforms as “essential infrastructures”.

162Initially, rights must be defined for the platforms as regards the conservation, consultation and sharing of data between people and organisations involved in public science before the data are disseminated to all users and beneficiaries.

163Françoise Curtit points out that her work on open science enabled her to discover the wide range of uses among different communities and the impact this has on the associated legal issues. A “wide-scope” model might not suit all disciplines. The variety of uses must also be taken into account when defining platform rights.

164The DIST and the Cabinet Alain Bensoussan propose to base the model on the generally accepted practices of each scientific community in order to institutionalise the uses. This first requires compiling the practices of the different communities.

165Paul-Antoine Hervieux and Françoise Curtit agree that this process is essential. The White Paper must take practices into account in order to define common orientations without dividing communities. The White Paper will define common guidelines that take different practices into account.

French Digital Council: Benoît Thieulin and Yann Bonnet, 1 September 2015

Meeting with the DIST and the Cabinet Alain Bensoussan within the framework of the drafting of the White Paper on Digital Science and Law

The process:

Drafting of a White Paper (notably) with a view to contributing to the Digital Republic Bill.

● Project led by the CNRS in conjunction with the Cabinet Alain Bensoussan,
● in response to the legal issues raised by the
ISTEX Investments for the Future project (, which involves two actions:

○ a vast programme concerning the acquisition of scientific resources, in the form of national licences;
the creation of a digital library with remote access for all members of higher education and research establishments.

● This project raises many legal questions due to its potential in terms of text and data mining
, interdisciplinarity, content aggregation and its aim of making its databases explorable.

● The White Paper aims to
propose a legal framework for scientific data in order to respond to the concerns of scientific communities (data publication and uses, law governing science platforms, text mining, etc.) and thereby contribute to the French Digital Republic Bill.

● The Cabinet Alain Bensoussan and the DIST are establishing working groups and conducting hearings to gather information about the community’s uses and the current state of the art regarding exploitation practices in order to establish a matrix concerning the relevance of analyses and practices, which will be compared to the current normative framework in order to assess discrepancies with the law and make proposals for the Bill.

Responses to the interview guide


Digital technologies are profoundly transforming the modes of production and dissemination of scientific results: data, publications and analyses are now accessible on various platforms. This availability of scientific material contains a potential for knowledge exploitation and sharing for which the law must be able to define the conditions, terms and limits.

166The problem today is that there is no specific legal status concerning scientific information.

167Scientific research studies and work are considered as written works, subject to copyright law.

  • To develop their careers and reputation, researchers must publish the results of their work in certain journals; to do this, they must transfer all their rights to the publishers.

168The publisher thus has the sole right to exploit, reproduce and disseminate the article.

169This means that research organisations and the scientific community cannot have access to the study/article unless they conclude an agreement with the publisher.

  • In addition to transferring all their rights, the research organisations must pay two types of costs:

    • upstream publishing costs;

    • downstream fees for consulting documents: higher education and research institutions spend more than €80 million a year to gain access to electronic resources. Access fees have also continually increased: 7% a year over the past 10 years.

170This represents a burden for public finances and hinders the productivity of public research, which must already cope with intense international competition.

  • As a result, there is a major imbalance between researchers and publishers, which has been aggravated over the last few years by the emergence of oligopolies in the scientific publishing sector (Elsevier, Springer, Wiley, etc.).

171With this in mind, we must assert the right to open access and open science, in order to respond to the concerns of the scientific community.

Definition of open access. Open access (OA) publications refer to articles that are accessible in digital format and can be read free of charge via the Internet. Open access allows readers to read, download, copy, distribute, print, search or create a link to the full text of an article, to index it, recover it for computer processing and use it for any other legal use with no financial, legal or technical obstacles, while fully complying with copyright.


  • reduce the burden of costs of digital journals in the budgets of public establishments;

  • facilitate access to scientific knowledge for the research community and civil society;

  • provide companies with broader access to the results of scientific research, notably small and medium-sized companies that could thus improve their innovation capabilities.

172The European Commission has invited Member States to “[d]efine clear policies for the dissemination of and open access to scientific publications resulting from publicly funded research” (2012).

CNNum proposals on open access

  • The conservation and dissemination of research results are public service missions. A movement supporting open publication already exists through open archive warehouses belonging to universities and organisations, or via the HAL platform.

  • Today, a specific legal framework would encourage this movement initiated by members of the scientific community.

    1. Recognise the right to secondary exploitation, as under German law. The author’s version deposited in an institutional archive remains in open access, whatever publishing course is later taken as regards the work.

      • 1 Higher Education Funding Council.

      Provide open access to scientific publications financed through public funds, after a short embargo period allowing for the publisher’s commercial activity, either in open journals or in an institutional repository (as under German and Italian law). This obligation should not lie with the researchers but with the research organisations.
      Care must be taken not to create a simple option to publish these documents in open access, which would render the rights ineffective (as is the case in Germany). It must be an obligation.
      The UK example is highly instructive. The HEFC,1 the organisation responsible for sharing out the overall sum of money allocated by the state between the different higher education establishments, has announced that any publication that is not available in open access as of next year will not be taken into account when assessing the activity of the establishments.

    2. Encourage researchers to give open access to raw, anonymous research data as long as this does not involve any issues regarding ethics or personal data.

CNRS’s proposals: Conditions, terms and limits to the sharing of scientific information


The position of the CNRS: The main conditions for free access to scientific results are the abolition of limits that may be introduced by editorial legislation (publication rights, copyright), with a view to the exploration of digital corpora of publications or data.

Question 1: What is your opinion on the necessary adaptations to publication rights (publishers and/or authors) and the exploration of corpora (text & data mining techniques, APIs, etc.)?


Definition of text and data mining. Text and data mining refers to various extraction and analysis tools that allow automated exploration of digital content, which can include text, data, sound, images or other elements, or a combination of these elements, in order to find new knowledge or ideas.

173Text and data mining techniques can help boost French research in the age of Big Data.

  • The MIT has qualified these techniques as one of the 10 emerging technologies that “will change the world” in the twenty-first century. Examples of applications:

    • The Text2Genome project made it possible to map the human genome by automatically compiling 3 million publications.

    • As regards the press, data mining is undoubtedly one of the future business models as regards information.

    • TDM is what enables Amazon to generate 20% of its turnover.

174TDM is not in itself a new activity.

  • It just involves reading and extracting information and meaning from documents. It is not really so different from gathering information manually, which has been the way of research since the birth of science.

175However, TDM often requires the creation of copies and content storage, which constitutes reproduction in terms of copyright law. Carrying out computerised processing of content repositories, whatever its nature (text, data, static or moving images, music, sounds, etc.) means taking possession of the content and storing it in order to search through it, or carrying out substantial data extraction, as regards the rights of database producers.

176The current normative framework does not allow TDM to fulfil its potential:

  • The individual management of rights is not adapted to such a large mass of data. In view of the many different sources and the large volumes processed, it would be impossible for the data miner to seek authorisations one by one.

  • The contractual solution is not satisfactory. The agreements proposed by scientific publishers severely restrict the uses that are authorised and sometimes set significant constraints and prohibitions:

    • Researchers are required to declare their research, which goes against all scientific ethical codes (research confidentiality, etc.).

    • Researchers are dependent on scientific publishers, which propose platforms with extremely limited services that are criticised by researchers. Contractual agreements make it difficult, if not impossible, to search between different repositories and datasets.

Proposals of the CNNum concerning text and data mining

177Introduce an exception to copyright law, without any compensation, providing for the right to use the information for TDM purposes within the framework of research.

178 This does not go against the exceptions established by the EU concerning copyright. The UK has such an exception through its “fair dealing” principle.


The position of the CNRS: The terms for sharing scientific results must, in the digital age, assimilate new constraints: the sharing of results between actors in public research on the one hand and between users and beneficiaries of public science on the other.

Question 2: How should the line be drawn between legitimate appropriation and misappropriation of results available on a public science platform, and how should these results be protected?

Position of the CNNum concerning the relationship between open resources and appropriation:

  • The benefits of open dissemination are currently underexploited by society as a whole. Often, it is the largest players, notably well-established web platforms, which seize these benefits by combining the commons with their own resources; thus, there is a real risk of predation.

  • The response to this risk cannot be to turn back and abandon this movement in favour of openness. The aim must be to help a large number of companies, associations, public organisations, researchers, media, etc. to develop their capacity to contribute to and participate in the commons, and particularly to use the resources.

179Possible use of non-commercial or share-alike licences: there is a whole range of contractual solutions available for managing the reuse of research results. Creative commons has developed solutions that are both easy to understand and machine-readable.

180We must nonetheless consider the objective of public research: is it not publicly funded partly in order to drive the economy and society as a whole? An increased number of licences may act as an obstacle to data exploitation and complicate the use of platforms. Ultimately, it is more of a political choice than a technical one as regards the public research model we wish to defend.


The position of the CNRS: Science platforms today contain STI whose form, content and legal status are very heterogeneous. This lack of uniformity impedes the visibility of science platforms. In the same way that there is today a notion of general interest data (a recent choice of the Minister of the Digital Economy), consideration should be given to the designation of public science platforms and, possibly, to a specific legal regime.

Question 3: Should we be moving towards a designation of “essential infrastructure” for major upstream research platforms, in cases where these platforms occupy a unique and irreplaceable function?

  • Platforms are becoming the gatekeepers of information. We need to rethink the concepts of competition law as regards “essential infrastructures” to impose an obligation of open access on these players.

  • This clearly raises the issue of independence. Do we want a private operator to control the access rules to public research knowledge and to make money from this access? I do not believe so.

181 This is the very idea of the commons.

  • The importance of free access.

For knowledge and access to knowledge as a commons, please see:
Charlotte Hess and Elinor Ostrom (eds.), Understanding Knowledge as a Commons, MIT Press, Dec. 2006.
Gaëlle Krikorian and Amy Kapczynski (eds.), Access to Knowledge in the Age of Intellectual Property, Zone Books, 2010.

ISTEX Committee, Laurent Schmitt, Jean-Marie Pierrel and Grégory Colcanap, 24 September 2015


For the ISTEX Executive Committee:

  • For the Couperin Consortium: Grégory Colcanap, Coordinator, and Monique Joly, Coordinator of the Studies and Forecasting Department

  • For INIST: Laurent Schmitt, Head of the Projects and Innovation Department, standing in for Raymond Bérard on the Executive Committee for the duration of his absence

  • For Lorraine University: Jean-Marie Pierrel, Professor

For the Ministry of National Education, Higher Education and Research:

  • Marie-Pascale Lizée, Scientific and Technical Information and Documentary Networks Department (DISTRD)

For the Cabinet Alain Bensoussan:

  • Alain Bensoussan, Barrister, specialised in law relating to advanced technologies

  • Laurence Tellier-Loniewski, Barrister, Director of the Intellectual Property Unit

  • Sarah Lenoir, Barrister, Intellectual Property Unit

For the DIST:

  • Renaud Fabre, Director

  • Laurence El Khouri, Deputy Director

  • Charlotte Autard, manager in charge of the ISTEX Investments for the Future project

Overview of the approach

182As mentioned during the ISTEX legal seminar of July 2014 and in the Executive Committee meeting, the Cabinet Alain Bensoussan, in conjunction with the DIST and the CNRS Legal Affairs Department, has undertaken to draft a White Paper entitled Open Science in a Digital Republic, in response to issues raised by the ISTEX project.

183This approach is based on the issues raised at the ISTEX seminar in July 2014 on legal security, on the themes: objectives and management of databases; text and data mining (TDM); interdisciplinarity; and content aggregation.

184These themes are of course included in the Digital Republic Bill drawn up by Axelle Lemaire, Secretary of State for Digital Affairs.

185The purpose of the White Paper is to contribute to an examination of the legal framework for scientific data in order to address the concerns of the scientific communities (data publication and uses, laws governing science platforms, TDM, etc.) and to participate in drafting the Digital Republic Bill by making useful proposals.

186The approach proposed by the Cabinet Alain Bensoussan is initially to conduct hearings in order to survey the practices of the scientific community and learn how STI is currently exploited.

187By surveying and compiling these current practices it will be possible to draw up a matrix of analyses and practices to be set against the current normative framework, in order to assess discrepancies and develop proposals, which will be submitted to all those involved in drafting this White Paper.

TDM and the right of observation

188The experience gained during the ISTEX project shows that it is more difficult to set up tools and rights for TDM in France in the current legal framework and under the current relationship with publishers.

189Grégory Colcanap considers it indispensable to obtain an exception to copyright, as is the case in the UK, to allow TDM practices to develop in France, together with a right to read.

190In France, publishing is mostly considered to be a cultural issue: an exception that had an adverse effect on publishing would be seen as an attack on French culture.

191Jean-Marie Pierrel recalls that TDM is merely automated reading and therefore basically no different to normal reading, except that it is done by a machine.

192The right to perform TDM is also the right to produce research data that become accessible to the entire community.

193Beyond the possibility of TDM, it is necessary to ensure the preservation of intermediate copies, i.e. annotated and modified documents. At present, some laboratories are already performing this preservation work but must now develop the tools for disseminating and sharing these data, as well as metadata.

194Depending on the disciplines and practices, metadata can be divided into two categories:

  • information associated with the document (metadata);

  • information associated with the way the document is used (usage data).

195User-generated content (UGC), a by-product of the use of article repositories, which is useful and recognised within the industry, must henceforth be included in these discussions in order to assess the added value of these usage data and determine how best to incorporate them in the assessment of scientific work.

196Renaud Fabre indicates that TDM is widely known and recognised as a practice, and has inspired considerable writing and led to strong positions being taken by such organisations as the ADBU, academic libraries, and all the directors and STI associations of research institutions.

197In the framework of the consultation on the Digital Bill, TDM as a tool for investigating the immense quantities of data and publications via electronic processing is a crucial issue for the future of research. Most of the major countries in research (Germany, Canada, United States, United Kingdom, etc.) have adopted such legal provisions: France cannot impose measures that set it apart from the international scientific community.

198Jean-Marie Pierrel indicates that researchers acquire TDM tools to search through texts or datasets in order to explain phenomena in extreme detail and to be able to use electronic processes to bring them to light.

199Researchers now need to be able to:

  • acquire primary data considering current market conditions;

  • access these primary data for unlimited observation.

200TDM must therefore involve not just a right to read but a right to observe. TDM is a right of observation of scientific objects, indispensable for science. If scientists have a universal and fundamental right of observation in the interests of scientific progress, then an exception to TDM is no longer necessary because TDM becomes a fundamental right.

201The right to observe scientific publications is necessary in order to be able to synthesise a multitude of observations across common areas. IT is just a specific device for observing data.

202The monopoly held by publishers aims to limit this right of observation of digital scientific databases by imposing the use of their own TDM tools and by controlling their use (APIs, limiting searches to a certain number of words, etc.).

203When acquiring the right to read content, this should not include any limitation on the right to observe.

Critical infrastructures and trusted third parties

204As TDM rights are recognised in ISTEX, this platform could be used as a model of a trusted third party for the depositing/archiving of intermediate copies.

205ISTEX could also be recognised as “essential infrastructure”.

206Other organisations, such as the French National Library (Bibliothèque Nationale Française, BNF), could potentially act as trusted third parties.

Embargo period and deposition

207Jean-Marie Pierrel considers that ideally there should not be any embargo period on scientific publications but only on scientific data.

208This is because an embargo period enables teams to benefit from the full use of the scientific data throughout the entirety of a research project or a thesis.

209In addition, the metadata must be accessible from the outset in order to ensure the sharing of information in the framework of ongoing projects in order to avoid duplicating costs, for example in projects using the same set of corpora.

210Marie-Pascale Lizée points out that the embargo period recommended by the European Commission (6 months after the date of first publication and 12 months for the HSS) is supported by the MENESR, but that opposition from publishers in the framework of the Digital Bill has led to a retreat in the Bill, with a proposal for an embargo of 12 months, and 24 months for the HSS.

211Jean-Marie Pierrel states that for some areas of research that require an embargo period on data of two to four years this should be possible (if justified).

212The 24-month embargo period for the HSS clearly expresses the fears of French publishers, but mainly concerns books and not articles. An embargo period on articles by the publishers of scientific journals in HSS would not affect them. French publishers must be educated in this respect to defuse the situation.

213Grégory Colcanap suggests that, concerning research data that are to be made available, some provision could be introduced in the funding mechanism stating that when data are produced in the framework of a project funded by a public body they must become public at the end of the project (or of an embargo period).

214The embargo period on articles by the publishers of scientific journals must be removed, but not that on books.

ABES hearing, Jérôme Kalfon, 5 October 2015


For the ABES:

  • Jérôme Kalfon, Director of the ABES

  • For the Cabinet Alain Bensoussan:

  • Alain Bensoussan, Barrister, specialised in law relating to advanced technologies

  • Laurence Tellier-Loniewski, Barrister, Director of the Intellectual Property Unit

  • Sarah Lenoir, Barrister, Intellectual Property Unit

For the DIST:

  • Laurence El Khouri, Deputy Director

Overview of the approach

215As mentioned during the ISTEX legal seminar of July 2014 and in the Executive Committee meeting, the Cabinet Alain Bensoussan, in collaboration with the DIST and the CNRS Legal Affairs Department, undertakes to draft a White Paper entitled Open Science in a Digital Republic, in response to issues raised by the ISTEX project.

216This approach is based on the issues raised at the ISTEX seminar in July 2014 on legal security, on the themes: objectives and management of databases; text and data mining (TDM); interdisciplinarity; and content aggregation.

217These themes are of course included in the Digital Republic Bill drawn up by Axelle Lemaire, Secretary of State for Digital Affairs.

218The purpose of the White Paper is to contribute to an examination of the legal framework for scientific data in order to address the concerns of the scientific communities (data publication and uses, the law governing science platforms, TDM, etc.) and to participate in drafting the Digital Republic Bill by making useful proposals.

219The approach proposed by the Cabinet Alain Bensoussan is initially to conduct hearings in order to survey the practices of the scientific community and learn how STI is currently exploited.

220The objective is to define the status that the scientific communities wish to see attributed to open science and, more particularly, what status should be attributed to data, metadata, articles or scientific results.

Data and publications: The ideal solution

221For Jérôme Kalfon the ideal solution for opening up access to data and publications is as follows:

222Above all it is necessary to avoid appropriation.

223For data: making data open is the basic principle, which must include justifiable (and usually temporary) exceptions. Precautions should be taken concerning the opening up of data, by defining the ethical requirements and the rules concerning confidentiality. The barriers to appropriation must be:

  • ethical security;

  • confidentiality;

  • the definition of a temporary monopoly and an embargo period on certain data, with the period varying on a case-by-case basis.

224While scientific publishing used to be only moderately profitable, the strategy of the major publishing groups has made it one of the most profitable sectors by manipulating two levers:

  • scientific publishing is a vital and strategic area for any academic structure, whether for researchers or for the institution or institutions on which they depend: as each new product is irreplaceable and exclusive by nature, the institution will be forced to accept any rise in the price of publishing, even if it is unrelated to changes in the costs of production;

  • the technological transition has upset the way things were traditionally organised and encouraged the concentration of actors to the extent that there is no alternative but to use them.

225For publications: the very purpose of a publication is to make content publicly available, and therefore no filter or embargo period should prevent or slow down publication. Publication should not be subject to appropriation of any kind. The bulk of the economic investment in publishing (writing, revising, basic editorial tasks, etc.) has always been provided by the scientific community. The publisher’s role – marginal if compared with the work involved in literary or artistic publishing – is limited to type-setting, layout, printing and distribution (managing subscriptions).

226Distribution mostly meant exchanging publications and distributing reprints. Publishers were essentially service providers, and a share of the potential profits from paying subscriptions went to the learned societies that owned the journals. The signing of contracts between authors and publishers is a recent development and has become more widespread with digital distribution.

227If it were necessary to invent a system of scientific publication today, we would never choose a system as costly, slow, inefficient and complex as the one in force now. At a time when rarity (physical limits to the number of copies and their distribution) is giving way to natural abundance, it is absurd and paradoxical that we should continue to create an artificial shortage.

228This anachronism has lasted for too long. It is based on an economic model involving the transfer of exclusive rights by the author to the publisher (a recent phenomenon that developed at the end of the 1990s – few written contracts between authors and publishers existed before that time).

229Consequently, publications in the world of science must be completely open, with no embargo period, subject only to the right of authorship.

230However, if “we leave behind wishful thinking and return to reality”, we need to define the conditions for a transition to a model that meets the requirements for building know-how and knowledge.

231A temporary embargo period could be established to cover a transition period, without losing sight of the ultimate objective of making publications available immediately. Embargo periods should be kept as short as possible and through negotiations with the publishers may be set at a maximum of 6 months, and 12 months for the HSS.

Remuneration for publishers

232Scientific publishing needs to find a new economic balance.

233Publishers must be paid for the work they perform, for the value they add, on a “jobbing” basis.

234Publishing is financed by the organisations that fund research. The contractual relationship between the researcher (the community) and the publisher must be translated by remuneration for a delivery of service. This service must be independent of copyright.

235Literary and artistic property (LAP) must be kept distinct from rights regarding science:

  • on the one hand, academic authors receive no remuneration from publishing (it is now extremely rare for authors to receive royalties – on the contrary, through the APC, they are now increasingly likely to have to pay to be published). Authors are paid for their academic activity, of which publication is an integral part.

  • on the other hand, the LAP system protects the authors of literary and artistic creations and provides them with remuneration. This traditional arrangement must not apply to science.

236For scientific publishing, there must be transparency regarding the added value for which the publisher receives remuneration.

237There is nothing to prevent scientists from publishing some parts of their work under LAP and other parts under open access.

238There are several legal considerations surrounding intellectual property today, which vary depending on the media and the type of product (audio, video, music, patents, trademarks, etc.). The failure to distinguish scientific publications from literary and artistic property is currently a problem for science, but the scientific publishing model is bound to evolve in a way consistent with its requirements. If no distinction is made, it is likely to be the LAP that comes under pressure.

Defining rights specifically for science

239It is therefore necessary to define rights specific to science (as there is for the filing of patents, for databases, etc.).

240Why do we need a specific law? Because:

  • science is a very special form of “trade”;

  • research activity, of which publishing is an integral part, enjoys specific tax regimes (tax deductibility) in almost all countries;

  • science is a factor of collective enrichment and development: scientific data, especially publications, must be treated as open data or content;

  • the use of this open content contributes to economic development;

  • the creation of wealth associated with the opening up of such content is much greater than that induced by their appropriation;

  • science is a particular field that follows a specific approach: researchers need to have access to and discuss texts and the results of research, and to repeat experiments (which in no way precludes filing patents);

  • copyright includes a right of withdrawal, which is incompatible with the ethics of publishing and scientific debate.

Distinction between Open Access and Open Process

241Just because a publication is in open access, this does not mean that there must be a transfer of ownership from the author to the publisher.

242Open access means giving access to the publication on the publisher’s website. It consists of a right to read on the publisher’s media. In a scientific article:

  • the idea must be entirely free to read;

  • the scientific article can be read on the publisher’s website (place chosen by the publisher for open access). Publishers restrict the exchange of ideas to a right to read.

243Open access on the publisher’s website must be distinct from deposition in an open archive. This deposition is based on the concept of filing the final version submitted to the publisher (and what then becomes of any differences with the final published version?).

244To overcome the problems related to intellectual property, the following proposals are made:

  • Establish the principle that publishers must deposit works systematically in an open archive (with remuneration for the useful work performed by the publisher); this open archive system is a necessary transition but must remain transitional.

  • Science as a set of specific rules for the development of knowledge should be devoid of authorship rights; only the moral right must be preserved.

  • STI must be defined as commons by its very nature.

The definition of a new ecosystem

245The following could define a new model:

  • publishers would be service providers: their services would include organising the peer review, labelling, assessment, and release for free access and free processing; the publisher would provide different services to the different communities;

  • the publisher would be paid for this service;

  • the final service proposed would be unlimited access and reuse of publications in any space, public or private, and including for commercial purposes: this would be “Gold Open Access” but without reversibility and without transfer of authorship rights;

  • the financing of the dissemination of knowledge would be integrated from the outset in the financing of the academic activity (from a macroscopic viewpoint, this is a zero-sum game: the sums spent on purchasing publications – mainly by libraries – finance the act of publishing, which is an integral part of an academic research project);

  • scientific publications would be copyright-free, without ownership rights; they would be a different kind of object: a “commons”.

246This system keeps publishers as professional third-party service providers within the dissemination chain of STI. Publishers cease to hold ownership rights over the content they help disseminate. The model is cleansed of any appropriation for the sole benefit of one of the actors, who previously gained economic and technical control of the entire value chain.

247All other systems are compromises. Embargo systems, hybrid journals, and recourse to TDM under contractual rules must be purely transitional solutions.

CCSD, Claude Kirchner, 15 October 2015


For the CCSD Steering Committee:

  • Claude Kirchner, current President of the CCSD Steering Committee, Adviser to the President of INRIA and Senior Researcher

For the Cabinet Alain Bensoussan:

  • Alain Bensoussan, Barrister, specialised in law relating to advanced technologies

  • Laurence Tellier-Loniewski, Barrister, Director of the Intellectual Property Unit

  • Sarah Lenoir, Barrister, Intellectual Property Unit

For the DIST:

  • Renaud Fabre, Director

  • Charlotte Autard, manager in charge of the ISTEX Investments for the Future project

Overview of the approach

248The idea by the CNRS to draft a White Paper entitled “Open Science in a Digital Republic”, in collaboration with the Cabinet Alain Bensoussan, is a response to the issues raised by the legal framework covering the ISTEX Investments for the Future project ANR-10-IDEX-0004-02 (, which raises many legal issues by its potential in terms of TDM, interdisciplinarity and content aggregation and its aim to make its databases searchable.

249The new law on science has a specific exception concerning copyright: the legal status of scientific results and notably data and metadata.

250The purpose of the White Paper is to propose a legal framework for scientific data in order to address the concerns of the scientific communities (data publication and uses, the law governing science platforms, TDM, etc.) and to participate in drafting the Digital Republic Bill by making useful proposals. Science has not waited for a Digital Act and is proceeding with practices that now need to be formalised.

251The Cabinet Alain Bensoussan will initially organise working groups and conduct hearings in order to survey the practices of the scientific community and learn how STI is currently exploited. The close association with universities and other institutions will make it possible to hear key witnesses and multiple views on the changes that are being prepared in the framework of the Digital Republic Bill.

252Surveying these practices and the state of the art will enable the CNRS to draw up a matrix concerning the relevance of analyses and practices which will be set against the current normative framework in order to assess discrepancies and develop proposals.

253Contributions to this White Paper will be anonymous and the people interviewed will have total control over whether their names are quoted or not.

254The DIST and the Cabinet Alain Bensoussan will re-contact the people interviewed and will submit an initial text to the University Presidents for approval.

Public consultation on the Digital Republic Bill

255In the framework of the public consultation on the Digital Republic Bill, Claude Kirchner met with the Secretary of State for Digital Affairs and the Director of her Cabinet and informed the representatives of the MENESR about the following points in particular:

  • All scientific data must remain under the control of scientists.
    Scientific data include texts, articles, webpages, calculation data, lab books, programmes, etc.
    But control does not mean possession. The objective is not to own the data but to have control in the sense of being able to use them in any way. Scientists must be able to access the complete text and the data in its entirety, so as to be able to read and reuse it, repeat and reproduce experiments, run programmes and be able to reference all or part of the appropriate data, while for all these actions complying with the ethical rules that apply to the areas of science concerned.

  • The services concerning data must be open to competition: there must therefore be complete access for TDM, in particular.
    For example, the CCSD is considering setting up services for data mining and analysis. Researchers must be able to access, and especially to verify data, whereas today the private sector withholds the information necessary for performing such verification.

256Opening suitable services would give the scientific communities access to the best possible tools and enable them to avoid losing their scientific sovereignty: the ability for a given discipline, or laboratory, or country to develop the best science at the international level.

257The competing interests of publishers in the management of scientific data might not be sufficient to force them to deliver all the necessary data to researchers.

258However, being a researcher means reproducing experiments, verifying data and the way the data were exploited.

259Today the possibility of publishers biasing search queries is plain to see for researchers and computer experts in algorithms and there is an awareness that while some of the strategies applied to algorithms can be shown to be fair, others may be biased. The recent case of Volkswagen has shown how important close supervision of access to data and algorithms can be.

260There is no proof at present that publishers knowingly bias access to their data, but the possibility of verification would remove any doubt, now or in the future.

261Article 4 of the Digital Republic Bill on the “Creation of a public service concerning data (to guarantee the quality of public reference data)” provides for this right of evidence, which is so vital for scientific publishing.

262The right of transparency expressed in Article 4 must be extended to include the queries themselves and the information revealed by the queries used and the work being done by those conducting the searches.

263The queries are themselves a form of data and must be accessible when they relate to scientific objectives. As long as ethical considerations are observed, the data from scientific discussions on social networks should also be accessible and free for dissemination.

264Concerning Article 9 of the Digital Republic Bill, the following remarks were made (the version of the article given below was the current one on the date of this interview):

I. When a scientific text arising from a research activity, which has been financed at least 50% by public funds, is published in a periodical, a publication appearing at least once a year, or in conference or symposia proceedings or compendia, its author, even in the event of exclusive transfer to a publisher, has the right to make available free of charge in digital form, subject to the rights of any co-authors, the latest version of his/her manuscript accepted by the publisher and excluding the formatting work which is the responsibility of the latter, at the end of a period of twelve months for the sciences, technology and medicine and twenty-four months for the human and social sciences, with effect from the date of first publication. This dissemination may not give rise to any commercial exploitation. The text of the Digital Republic Bill – Article 9.”

265This wording prompts the following comments.

266On the section “… even in the event of exclusive transfer to a publisher, …”, as Roberto Di Cosmo states in his comment on the proposal of Olivier Morin for “full and mandatory free access” in the framework of the public consultation on the Digital Republic Bill, if researchers do not transfer their rights to a publisher, they cannot sign the contracts currently written by publishers that impose the transfer of rights for the publication of articles in prestigious journals. It is therefore essential that the Act guarantees free dissemination, online, elsewhere and immediately, without contract assignment.

267... the latest version of his/her manuscript accepted by the publisher and excluding the formatting work which is the responsibility of the latter”: Claude Kirchner first insists on an important element of the process of publication in a journal. A scientific paper submitted for assessment by a journal’s editorial board has three main successive versions that it is important to identify correctly. The author’s initial version, also called the author-submitted version, is the version of the document as sent for assessment to the editorial board. The author’s accepted version is the one declared accepted by the editorial board after the peer review. The publisher’s version is the one that, based on the author’s accepted version, has been edited by the publishing house to improve the presentation and style, and put in the right format for the journal if the authors have not done this themselves. This is the version that will be published in the journal.

268Since only the peer-reviewers have worked on the author’s accepted version, it is inappropriate to apply an embargo period: the draft article is therefore incorrect here. The version on which an embargo period could be applied is the publisher’s version, that is to say the one that the publishing house has helped to format.

269These reflections also led Claude Kirchner to point out that the current transformation that scientific communication is undergoing reveals three fundamental facts:

  • Publishing a text, a result, a study, or data involves making these items public by definition. Publishing may be preceded or followed by qualification and validation as explained below, but it is fundamental to note that today, anyone can publish. People can tweet, post texts on their blogs or their webpages, publish articles in a magazine, give papers at a conference, post opinions or texts on a scientific social network, etc.

  • Qualification: A document, a text or a dataset, whether published as explained above or not, can be reread to assess the quality, originality, the contribution compared to the state of the art, the referencing, the quality of writing, etc. This qualification can be carried out by a journal’s editorial board or by the programme committee of a conference, in which case we speak of peer-reviewing. But this qualification is also often carried out elsewhere: in correspondence exchanged between a small group of people, on scientific social networks, in discussion forums specific to a community or in workshops. It can also be done during the assessment of research proposals for the French National Research Agency (ANR), the European Research Council or Horizon 2020, for example. In the framework of citizen science, it may be the result of different processes, such as voting.

  • Validation or certification: Following a qualification process, a community of researchers such as, for example, the editorial board of a journal or the programme committee of a conference may decide to validate (or “certify”) the document or the dataset. In the case of a scientific text, this validation by an editorial board may result in the acceptance of the article in a journal. Because before the digital revolution it was difficult or impossible to publish outside a magazine, there is still confusion between validation and publication. It is vital to note that these two elements are fundamentally different.

270In this framework, quite independently of the consultation and following a long process of reflection that started at the very beginning of the 2000s, the President of INRIA very recently distributed a note that applies to all the scientists conducting their research in an INRIA project team. The note required that all works be published in full text in an open archive such as HAL and that for assessments of the Institute, its project teams and its researchers, only these publications will be taken into account. In the event that these works are also made public in a journal or a conference, the note requested that the version deposited in HAL is either the initial version of the article or (not exclusively) the author’s accepted version. The note also stated that if a publisher seeks to oppose this deposit on HAL, challenging for example the free posting online of the author’s accepted version, INRIA undertakes to assume this responsibility.

271Availability: the phrase “This dissemination may not give rise to any commercial exploitation” inspires the following comment. The content of a scientific text is of course potentially a source of innovations with considerable commercial benefits. The transfer of scientific progress, and therefore of the texts that describe it, is one of the fundamental missions of scientists in research organisations and universities. To prohibit the commercial exploitation of the contents of a scientific article by its authors and their employers would therefore be contrary to the fundamental missions of schools, research institutes and universities.

272II. – The provisions of this article are public policy and any clause to the contrary is deemed to be unwritten. They shall not apply to contracts in progress.” At the present time, the current contracts with the major publishers are signed for periods up to 2018. It would therefore not be possible to apply any new Act until the end of this period.

273The Act must formally state that it applies to contracts currently in force or this clause must be removed to leave the door open.

Legitimate appropriation and misappropriation

274Claude Kirchner stated that all the scientific data, at least before their publisher’s version, must be made available to all comers, whether public or private.

275During the previous hearings a consensus could be seen emerging on free access to data, but not concerning the embargo period. This can differ depending on the practices of each discipline and the researchers’ need to be able to exploit their research prior to its dissemination and sharing.

276According to Claude Kirchner, today the essential element when a result is made public is to have determined beforehand whether or not it should be protected.

277Before a theorem, a process, an article, data, a programme or an algorithm is posted online, INRIA takes careful heed of the willingness or otherwise of the researcher and the co-authors to publish the material in question.

278In some cases:

  • the Institute may issue an unfavourable opinion concerning publication but the author has the final decision and all the elements necessary to make that decision;

  • the Institute may prohibit dissemination if the content fails to comply with its ethical rules. There are in fact a number of publications or programmes that cannot be made public directly, such as for example a programme capable of breaking encryption or hacking into bank accounts.

279The different entities that oversee research activity should be organised in such a way as to allow the verification of publications in OA (compliance with ethics, possibility of exploitation) and to inform researchers prior to deposition in an open archive.

Services and innovation

280In order to ensure the control of data, the CCSD via the platform HAL wants to create services of international quality. If the government supports this Bill and accepts this text for Article 9 on open access, “A shorter embargo period, no hindrance to TDM (text and data mining) and no prohibition of commercial exploitation”, as proposed by the DIST, it will also be necessary to set up public sector services capable of competing with commercial offers.

281These public services could, for example, take the form of a recognition of some science platforms as “essential infrastructures”.


1 Higher Education Funding Council.


Open access


Volume papier