This case study analyses the process of translation of a popular science (picture)book that originated from the Internet comic strip xkcd. It explores the obstacles resulting from the text-image interplay. At the macroscale, while such institutions in Poland as the Book Institute or translator associations do develop standards and provide information on the book market and good (or actual) practices, they never explicitly mention comic books – the closest one can find is “illustrated books” or “others”. Additionally, popular science literature – uninteresting, one would say – is much less discussed than artistic translation (with due allowance for comic books and graphic novels) despite having a tradition of using words and images together. Thing Explainer does seem to use “other” translatory techniques: firstly, because the author decided to use only one thousand most common English words (a semantic dominant to be retained in Polish); and secondly, because the illustrations – from diagrammatic to extremely detailed – are an indispensable, though variably integrated with the verbal, medium of knowledge transfer. This paper focuses on the second aspect. Specifically, it discusses: rigorous requirements for text volume and location (exacerbated by the said 1000-word list); technical issues (including choosing typefaces, formatting text, modifying graphics etc.); overlapping responsibilities of the editors, the translator and the DTP artist; and unusual text-image relations (e.g. the image helping a translator decode what he actually translates).