Seminar given in Nice, France, 19th November 2014 (in French). English abstract: The advent of things like digital content, the internet and the movements on open access and open science has profoundly changed the playing field in the dissemination of science. Nevertheless, the scientific article remains the dominating way to validate and to diffuse scientific endeavors, and the central components in evaluation of science are still the journal impact factor, and various forms of publication and citation statistics. This may be about to change. Traditionally, scientific interactions other than articles, such as discussions, debates, seminars and media coverage, have mostly been ephemeral in nature. This is no longer strictly true. Reports in newspapers, radio and TV leave searchable traces. Discussions about science are increasingly held on blogs and on social media. Research leaves imprint in legislation, patents, policy documents etc., and some scientific journals have initiated post-peer review. Technology for data mining of the internet now exist, which enables the assessment of what impact a certain research effort has had, beyond citations. There are already companies that do this, and their services are for example bought by some publishers. This is presented as a service to authors, who get information about the impact of their published work. There are also some funding agencies who increasingly ask for information of the impact of the research they fund, other than just citations. To a large extent, the developments in this field, as in scholarly publication in general, are driven by the demands in biomedicine and life science. This is where most of the funding is, and it is also where scientific interaction is most intense and immediate. However, the changes that may surface in the evaluation and dissemination of research will have an impact on all fields of science.