Tables for
Volume G
Definition and exchange of crystallographic data
Edited by S. R. Hall and B. McMahon

International Tables for Crystallography (2006). Vol. G, ch. 2.1, p. 13

Section 2.1.1. Introduction

S. R. Halla* and N. Spadaccinib

aSchool of Biomedical and Chemical Sciences, University of Western Australia, Crawley, Perth, WA 6009, Australia, and bSchool of Computer Science and Software Engineering, University of Western Australia, 35 Stirling Highway, Crawley, Perth, WA 6009, Australia
Correspondence e-mail:

2.1.1. Introduction

| top | pdf |

A human language, in all its forms, is spoken, written and comprehended according to grammatical principles that evolve continuously with the need to express efficiently new ideas and experiences. In this chapter we will describe how similar principles have been developed to construct, describe and understand scientific data, such as numbers and codified text. The efficient and flexible expression of data may be achieved using grammatical rules similar to those of spoken languages, though the precise and unambiguous rendition and communication of data must preclude those nuances, subtleties and individual interpretation so important to the spoken and graphical world of poetry, literature and art. It is important to record these aspects of human endeavour, but they are distinctly different from the objectives of science, where information must be expressed with maximum precision and efficiency.

Understanding any language involves two fundamental steps – identification of the individual elements of a language sequence, and comprehension of the sequence structure. In a spoken language these steps normally comprise the simultaneous recognition of individual words and the understanding of their grammatical context. Such a simple decoding process belies the enormous potential for complexity in spoken languages. Nevertheless, most `word sequences' are understood in this way and a similar approach may be used with non-textual data.

As discussed in Section 1.1.3[link] , until quite recently most approaches to storing scientific data electronically were based on fixed-format structures. These are simple to construct and easy to comprehend provided the data layout is fixed and widely understood. That is, items, as well as lists of items, are written in a fixed sequential order that is mutually agreed on by those writing and reading the data. However, the preordained nature of a fixed format, which intentionally prevents changes to the data structure in a file, also poses a serious limitation for many scientific applications. This is because the nature of data used in scientific disciplines, such as in crystallography, evolves continuously and requires recording processes that are extensible and adaptable to change.

A lack of ready extensibility in the expression of data is particularly problematical for long-term archiving. For example, the recovery of information stored in a fixed format may be impossible if the layout details are lost or altered with time. Less rigid fixed-format variants, using keywords to identify groups of data, can improve the flexibility of data recording but they often preclude the introduction of new kinds of data.

In the 1980s it was widely recognized across the sciences that more general, extensible and expressive approaches were needed for recording, transmitting and archiving electronic data. This led to the development of free-format file structures. These file structures are the basis for universal file formats intended to: (a) store all kinds of data; (b) be independent of computer hardware (i.e. portable); (c) be both machine-parsable and human-understandable; (d) adapt to future data evolution (i.e. be extensible and robust); and (e) facilitate data structures of any complexity (i.e. have rich syntax capabilities).

The extent to which these objectives can be met determines the universality of a data-storage approach. Several of these properties are difficult to achieve simultaneously, and the compromises that have been adopted have largely determined the success of universal data languages in different fields. The STAR File is a universal data language applicable to all scientific disciplines, and the Crystallographic Information File described in Chapter 2.2[link] of this volume is a specific instance of the STAR format that has been adopted by the crystallographic community.

to end of page
to top of page