One of Himmelmann’s primary goals in his 1998 paper was to argue for a strict division of documentation and description. Language documentation has since successfully developed to become a discipline in its own right. Nevertheless, the question concerning the interrelation of description (and thus analysis) and documentation remains a matter of controversy. This paper reflects on descriptive and documentary adequacy, focusing on two major issues. First, it addresses the question of how much analysis should enter into an adequate documentation of a language and, second, it discusses the role of language documentation and primary data in the replicability of linguistic analyses.