Free scientific data circulation within an open environment.

Last July, Laurent Romary, Research Director at INRIA, the French national institute in Computer Science, was appointed head of the Scientific Information Directorate, a newly created entity within the French National Center for Scientific Research (CNRS) designed to support the implementation of a coordinated policy for scientific and technical information.
Perspectives…

Question: Last July, you were appointed head of the Scientific Information Directorate. What are your major projects today?

Laurent Romary: When speaking of Scientific and Technical Information or STI, one must keep in mind that STI goes well beyond issues related to publications. The will to improve the dissemination of the bulk of the researchers’ output-publications and primary data (lab logbooks, statistical data, photographic or multimedia materials)– has been the driving force behind the creation of the Scientific Information Directorate. Now, we have to give ourselves the means to improve the circulation of all these data, not only upstream of publication as with ArXiv, but also in a synchronized manner, publications and primary data, as is the case in the Life Sciences field.

Q.: What about the primary and nontextual data that are essential to many disciplines today?

L. R.: As far as I am concerned, I see two directions. The first which consists in identifying the primary data and harvesting the metadata for all those materials, and the second which consists in increasing the standardized digitization of data and documents not yet available in digital format. It is along those lines that competency centers are being set up to provide all laboratories with the same access level to information, along with a strategy adapted to the various disciplines.

Q.: Can you tell us about the cooperation with other public scientific and technical research organizations and universities as far as archive input and access are concerned? Is interoperability the rule?

L. R.: Factual data are essentially subject-oriented and closely related to the concerned field. Therefore, they do not require the implementation of extensive cooperation between research organizations. This is different with publications or databases where, through a single national cooperation, we can have an overall vision. At this time, we are counting on political contributions rather than on technological developments to move toward a single publication database. The majority of CNRS laboratories are joint units working with other research or higher education organizations. Currently, we are mostly working with these players.
Today, we are getting close to an agreement with the French institutes for agricultural research (INRA), health and medical research (INSERM), research in computer science (INRIA) and French universities. During the next couple of years, we are going to experiment with the joint administrative and technical management of a publication archive.

Q.: Several studies conducted in Europe and in America have shown that researchers must be involved for open archives to develop. Are specific actions envisaged at the CNRS level?

L. R.: It is obvious that the researcher must be at the center of the archiving system. As far as the institutional archive is concerned, which will be based on HAL ^[1]The HAL “Hyper Article en Ligne” archive provides authors with an interface enabling them to deposit in the database of the CNRS Center for Direct Scientific Communication (CCSD/CNRS) manuscripts … Continue reading , researchers will have an input interface which they will have to complete when they deposit their articles. The major argument against this is that the additional administrative burden will hamper the archives’ development. It is out of the question to create a service dedicated to fuelling the archive and in which no researcher would be involved. Actually, all researchers already accomplish similar tasks at assessment time, when they answers calls for tenders or when they update their bibliography on their personal web pages. Several support measures should act as incentives: a complete range of tools and services will available to researchers to help them with their research. For example, they will no longer have to give lists of publications since these will be retrieved directly from the archive.

Q.: The British Research Councils recently proposed that mandatory archiving be part of public research funding. A proposal already existed in the OECD already issued recommendations in January 2004. Could such provisions be considered in France?

L. R.: Probably, but one step at a time. On the other hand it is not advisable to rush things by applying constraining measures without taking into account certain specificities. You cannot ask researchers to all of a sudden adopt self-archiving. When it comes to primary data, things are somewhat different. More pressure can be applied than with publications since no special precautions need be taken as these data fall outside of the publishing cycle. This is why the CNRS favors a cross between an open archive and an institutional archive as well as a strong partnership approach with all institutions in order to reach systematic deposit. Then, depending on disciplines and topics, data will be openly available or restricted for confidentiality reasons. The approach is changing but our objective remains similar to that of a Stevan Harnad, that is to create in the end a environment where data produced by research activities can freely circulate.

French version