One should begin by asking the fundamental question, ``Why use a standard format?'' What does one gain by using such a format over the special-purpose, machine-specific formats commonly used in the past?
Three major reasons come to mind. The first is portability; a standard format makes it easier to move datasets to the new computers one acquires over the years, as well as to machines at collaborating sites. The second reason to use a standard format is so that the contents of a strange file can more readily be understood. Such a file may not necessarily be from a foreign site; one's own datasets, inspected years after the original programmers and project scientists have passed into the dust of history, can seem very unfamiliar. The third reason is to avoid re-inventing a new format each time a new dataset appears: using a standard format allows greater re-use of software to read and write data and also permits the construction of a set of tools which can be applied to a wide variety of data. A standard format, if it is to be useful, must therefore possess certain qualities which fulfill these three basic needs: portability, understandability, and reusability.