Born in the 17th century, journal peer review is an extremely diverse technology, constantly torn between two often incompatible goals: the validation of manuscripts conceived as a collective industrial-like reproducible process performed to assert scientific statements, and the dissemination of articles considered as a means to spur scientific discussion, raising controversies, and civically challenging a state of knowledge. Such a situation is particularly conducive to clarifying the processes of valuation and evaluation in journal peer review. In this article, such processes (...) are considered as specific tests in order to emphasize the uncertain properties of pre-tests manuscripts. On the one hand, evaluation tests are examined at the core of the validation of manuscripts, such as defining the coordination of judging instances or controlling the modalities of inter-knowledge between reviewers and authors. They are also studied regarding the dissemination of articles, notably through the contemporary conception of a continuing evaluation test termed “post publication peer review”. On the other hand, valuation tests are both part of the validation of manuscripts, such as the weighting of different judgments of the same manuscript and the tensions that these hierarchies cause, and of the dissemination of articles, such as attention metrics recording the uses of articles. The conclusion sketches out how the articulation of these different tests has recently empowered readers as a new key judging instance for dissemination and for validation, potentially transforming the definition of peers, and thus the whole process of journal peer review. (shrink)