Data exists; it is as simple as notes taken from historical sources. We tend to call it “data” when we take our notes in a spreadsheet, but it is really any type of notes taken systematically from historical sources. Our spreadsheets are verbose and most “digital humanists” would want to “clean” them, but we cherish their dirtiness. It is a sign that the transformation from source to data (notes) happened through our own thought processes (not those of underlings paid to sweep the dirt under the rug) and retained the ambiguities of the sources. To make data better, and to have more colleagues crave it and fewer abhor it, we want to keep it as close as possible to the source, even though it will be dirtier, and more costly to produce in large quantities. But that’s fine, because we do not believe data is good only if it is really big. It is good if it is complicated, and thus rich in information, but still systematically acquired and noted in a structured way, so that we can simplify it in many different ways if we want to experiment with it. It is good for thinking, even, or especially, when it produces new questions rather than final answers. This was the manifesto in our title. To flesh out what we mean, this paper will illustrate three main points through cases from our research on apprenticeship in France.