Indeed, since August 2005, Ifremer co-published 990 articles referenced in the Web of Science®. 812 of these 990 publications, almost 82%, are freely available in Archimer.

• 31 articles were published by publishers who were not yet listed on the Sherpa/Romeo website (National Shellfisheries Association, International Phycological Society...) and needed to be contacted,

• 40 articles were published by publishers who forbade the registration of their publications in an Open Archive (ex: American Meteorological Society, ASLO, American Chemical Society ...),

• 177 articles were published by publishers who allowed self-archiving of their own PDF files,

• 742 articles were published by publishers who limited the right of self-archiving to the author’s final manuscript. The drafts of 613 of these 742 items were collected and recorded

This good result is linked to the involvement of Ifremer’s library in the operation of this archive. It is the library staff itself that ensures the preparation and registration of publications into Archimer:

• Each week, we search for the publications written or co-written by Ifremer in the Web of Science®,

• Then for each publication, we consult the publisher’s policy on the Sherpa/Romeo website. If the publisher’s policy is not specified on the Sherpa/Romeo website, or on its own website, we systematically contact the concerned publisher to ask permission to record the articles into Archimer,

• If the publisher allows the self-archiving of its own PDF files (ex: EDP Sciences, The Company of Biologists ...), we download ourselves the PDF file of the concerned article from the editor’s website and then record it into Archimer,

• If the publisher restricts the right of self-archiving to the author’s final manuscript of the publication, we contact the authors of the publication to request a version using some automated tools developed by Ifremer. If they can provide us with this version, we issue ourselves a PDF file from the files sent, and then register it into Archimer. In most cases, these publications are sent under multiple files (one file for the text and other files for tables and images). We merge all these files, we create a cover sheet with the references of the article, we use a minimalist layout of the text before turning it into PDF and optimizing it for a better visibility on the WEB (ex: http://www.ifremer.fr/docelec/doc/2008/publication-4501.pdf )…

As a consequence, we contact authors all through the week to collect their publication. A recovery system, also automatic, enables us to remind the authors that we expect their paper for Archimer. In most cases, we need to contact the authors 2 or 3 times (and sometimes more) before obtaining their publications. This harassment policy causes few calls from scientists fed by these reminders, but fortunately, this work is generally very well received.

Obviously this method is not perfect: it is costly in time (1 working day for 10 publications), fragile because of the financial difficulty to hire staff and the collection of documents based only on spontaneous deposit does not work (conference proceedings, internal contract reports ...).

But this method enables an important collection of publications referenced in databases. It helps to overcome :

• The authors’ lack of time,

• Their lack of immediate interest. If the promotion of free access is not necessary for physicists for whom ArXiv has become unavoidable, this is not the case for other fields like, for example, life sciences - in majority at Ifremer. For these scientists, tools like the Web of Science® or ScienceDirect®, Open Archives are not yet working tools,

• Some scientists’ (this is particularly true for people working on life sciences, in majority, at Ifremer) lack of computer knowledge. A scientist answered our request for his final manuscript saying: "The version can’t be released (heavy file, figures and text separated)",

• The ignorance of copyright rules

This method has also other advantages:

• Saved documents are optimized for better visibility on the WEB,

• These personalized contacts with authors enable us to increase their awareness of the practical aspects of the Open Access movement faster and especially to make them understand the importance of their final manuscript,

• The recording of an important number of documents will enable a fast collection of a critical amount of documents. This amount, will mechanically enable a better visibility of the project, and could quickly give us :
  • The recognition of Ifremer’s authors (more spontaneous deposition...)
  • The recognition of Ifremer’s leadership (more resources...)

Today, we hope to go beyond these 80%, continuing to educate Ifremer’s staff about the challenges of the Open Access movement. We are also trying to establish a systematic collection of publications from their journal acceptance to make sure we have all the elements necessary for an online publication. Although these deposits represent a minority, we noticed a growing participation to this scheme.

However, for the collection of publications, we think we have already reached a limit. ,The missing 20% are indeed mostly co-publications where the involvement of Ifremer is minor and anecdotal. Moreover, we will always be restricted by a small amount of articles published by publishers that prohibit self-archiving, or by the loss of some author’s versions... In the coming years, we hope, of course, continue the collection of a maximum of publications but also to increase our efforts to collect other types of documents, including different types of reports (internal contracts reports...) we are missing for the moment.

Please contact us if you want more information about this experience.