The general concepts and requirements of a universal exchange file are introduced in this chapter. The Self-defining Text Archive and Retrieval (STAR) File was designed to meet the requirements for a universal data-exchange mechanism, and it forms the basis of the format widely used in structural science to exchange and archive numerical and text data. It was published in 1991 and extended in 2012. The STAR File structure is simple. Discrete data values, always represented by character strings, are identified by distinguishable unique tags. A file is partitioned into one or more blocks of data. Within each such block, one or more subsidiary blocks (known as save frames) may encapsulate a set of data items. In each block, a given tag may appear once only, but may index a single data value (which may be simple, comprising a single text string, or complex, with a number of discrete elements) or a set of looped values. Loops of related items may be combined in tabular structures. Global blocks may occur that establish data values inherited by succeeding data blocks. This chapter describes the STAR File syntax and scoping rules in detail.