The partial structures approach in the philosophy of science, popularized by Bueno, French, Ladyman, and others, has always been advertised as allowing for a realistic formalization of scientific theories and models, their development, and their relation to observations and measurements. I argue that this claim has never been established because all alleged applications have relied on imprecise paraphrases in ordinary language, and especially an ambiguity in the term ‘relation’. Furthermore, the definitions of ‘partial structure’ and ‘partial isomorphism’ make an application of partial structures so unwieldy that any advantage over standard model theoretic structures is lost.