Philosophy of science in practice is philosophy of clutter, and not of theory. While the world of the theory is clean and clear, the world of research practice is messy and noisy. Most of the research time is spent on cleaning, filtering, and similar tidy-up activities. Clutter leads to errors. Hence, the more can be cleaned up, the more accurate the research results will be. But clutter is heterogeneous in the sense of its composition, idiosyncratic with respect to its environmental conditions, and sticky, that is, hard to separate from the object of study. Due to the nature of clutter, no theory can account completely for the practice of research. For the epistemological understanding of practice one has to study other documents, namely almanacs, dictionaries, guides, handbooks, instructions, reports, teaching materials, tutorials, and yearbooks. Although for the study of tacit knowledge ethnographic methods seems to be most appropriate, these latter documents provide rather detailed accounts of these idiosyncratic practices. An exemplary document is G. Girard (1990), ‘The washing and cleaning of kilogram prototypes at the BIPM.’ For the same reason as there is no theory of clutter, the treatment of clutter cannot be done by only mechanical procedures. Typical for this kind of documents is that they also instruct about essential non-mechanical activities as “rub fairly hard by hand”, or “give a few taps on the instrument.” Because there are no standards for how hard to rub or how many taps, these judgments are often based on visualizations. (I will not discuss the equally interesting judgments based on smelling, hearing, tasting and touching.) These judgments require training and accumulated experience with the specific practice. Philosophy of science in practice therefore is a philosophy of idiosyncrasy and trained senses.