Traditional practice has been to standardize on a single general-purpose programming language for large software projects (e.g. Java, C++). More recently though, DSLs have been seen as a way to program in a language tailored exactly to the needs of various parts of an application. The new criterion for choosing a language may be not what it can /do/, but what kinds of languages it can /support/. Using miniKanren as an example, we survey the more that 40 hosts that have “Kanren” implementations, and demonstrate what separates an adequate host from a great one.