Research assessment is going to play a new role in the governance of universities and research institutions. Evaluation of results is evolving from a simple tool for resource allocation towards policy design. In this respect "measuring" implies a different approach to quantitative aspects as well as to an estimation of qualitative criteria that are difficult to define. Bibliometrics became so popular, in spite of its limits, just offering a simple solution to complex problems. The theory behind it is not so robust but available results confirm this method as a reasonable trade off between costs and benefits. Indeed there are some fields of science where quantitative indicators are very difficult to apply due to the lack of databases and data, in few words the credibility of existing information. The University of Milan started in 2004 an institutional repository which contains the scientific production (bibliographic metadata and full-text if possible) of all researchers in all research fields. The present work describes the results of two assessment exercises performed on the basis of the data of this institutional repository. The results of the internal exercises are quite satisfying. Indeed they have emphasized some problems but they have already been solved. There is a problem of comparability at national level due to the lack of reliable databases that could be solved by a widespread adoption of institutional repositories.
Keywords: Open access; Institutional repositories; Humanities; Universities; Research assessment.