The most important component of the Crystallographic Information Framework is the controlled vocabulary that attaches specific, well defined meaning to each of the data names defined in the CIF dictionaries. These dictionary definitions are intended primarily for human comprehension and interpretation. However, each definition is accompanied by a set of machine-readable attributes describing type, permitted values or value ranges, aliases and relationships to other data items. A number of tools can extract such information from the CIF dictionaries and use it to assess the reasonableness of values that appear in data files, and in certain cases to evaliate missing values. This chapter describes approaches to such validation and in particular describes tools that can parse and execute methods written in the relational expression language dREL. It also provides a survey of crystallographic software packages where the semantic information about domain concepts is hand-coded and not directly reliant upon CIF dictionaries, but that nevertheless make full use of the CIF format for data input, output and transfer.