Computerised glow curve fitting methods have been used by physicists studying thermoluminescence (TL) mechanisms for at least two decades. Applications of computerised glow curve deconvolution (CGCD) to thermoluminescence dosimetry (TLD) began to appear in the scientific literature in the early 1980s, not far behind the introduction of personal computers into academia and industry. It was fairly obvious that the application of well known mathematical fitting routines to the glow curve of most common TLD materials might lead to important advances in TLD. Lucas pointed out some of the many potential dosimetric applications including: (i) measurement of neutron equivalent, (ii) fading corrections, (iii) determination of time from exposure, and (iv) detection of various types of system failure. Demonstrated improvements in precision of measurement and minimum measureable dose levels soon followed. The first generation of personal computers from Apple Inc. (Apple II, Apple IIe and Apple IIc) contained 128 kBytes of memory and, at best, a 1MHz clock. It was not at all clear at the outset, therefore, whether these computers possessed the speed and versatility required to handle TL analysis. Indeed, on these computers, the iteration time for the deconvolution of a glow curve consisting of seven glow peaks (21 variable parameters for the seven glow peaks and three variable parameters for the background) was approximately 15 min. With an Apple computer running overnight one was not usually successful in deconvoluting more than 6-7 glow curves. Unlike the situation in other spectrum unfolding applications (e.g. nuclear gamma ray spectra), the TL peak 'line shape', i.e. the shape of a single glow peak, is a priori unknown. The line shape is a complex, fairly intractable, mathematical entity. It is, furthermore, model dependent, usually varies from peak to peak in the glow curve, and can be dependent on a frustrating variety of both facility (readout), annealing, level of dose and radiation field parameters. This is both a curse and a blessing, mainly the former. On the one hand, the line shape can be used to characterise phenomenologically the underlying TL mechanism, i.e. one can do a sort of physics by studying the line shape and its dependence on various factors. On the other hand, the applications to dosimetry can get lost in insensate quibbling about these phenomenologically derived factors and their possible meaning. Another equally serious complication arises from the loss of universality so common to TLD. Nonetheless CGCD has emerged as an important and powerful tool in TLD applications.