Some hope for scientific information databanks in the social science?

Analysis of the problems met in the practice of bibliographic databanks in the Humanities and social science, and attempt to understand all the reasons, other than economic, of the low use in bibliographic databanks related to basic research. The use of the traditional documentation model, which does not meet the needs of the current research, seems to be the main reason. Directions for research are suggested so as to adjust information processing to the behaviour of the different types of users.