Open archives, the expectations of the scientific communities

<!--HTML--><p align="justify">Open archives (OA) started in physics more than 15 years ago with ArXiv, and have since played a more and more important role in the activity of the disciplin; actually, in many fields of physics, ArXiv has now become the major vector or scientific communication. We now have two communication channels in parallel, traditional scientific journals with peer review, and open archives, both with different functionalities and both indispensable. It it therefore interesting to try and transpose to other disciplins the scheme that has worked so well for physicists, which means that the reasons for the success of ArXiv should be analyzed. <br><br> Scientists do not care about the technicalities, and whether or not the OA is centralized, or distributed with a high level of interoperability. What they wish is to have one single interface where all the scientific information in their domain is available, with the same scientific classifications, etc.. In case of collaborations beween different institutions, they do not wish to have to deposit their texts several times in several repositories. Moreover, they wish to use OA that are relatively homogeneous concernig the content, and for instance do not want to use systems where the free access to the full text is not garanteed, or articles are mixed up with reports or raw data (as is seometimes the case in OAI repositories). Finally, long term stability is absolutely essential, if only to offer the possibility to quote the content of the OA in subsequent work. In this respect, we know that much of the material made available on personal sites and laboratory sites does not have the necessary longetivity to be really useable scientifically. <br><br> This analysis led the French CNRS to start the Hal project, a pluridisciplinary open archive strongly inspired by ArXiv, and directly connected to it. Hal actually automatically transfers data and documents to ArXiv for the relevant disciplins; similarly, it is connected to Pum Med and Pub Med Central for life sciences. Hal is customizable so that institutions can build their own portal within Hal, which then plays the role of an institutional archive (examples are INRIA, INSERM, ENS Lyon, and others). It includes a form of certification called "stamps". The rate of increase of the number of full text documents increases steadily, and now reaches around 1 500 per month, mostly due to self archiving from different communities (including humanities: archeololy, history, literature, etc.). A national agreement has been signed among most research French institutions to use Hal as a common archive. <br><br> <a href="http://cern.ch/oai5/presenters_profiles.html#LALOE" target="_blank">View Franck Laloe's biography</a> </p>