``For the highest payoff per dollar risked it should be remembered that communication system development is normally a ten year leadtime cycle process. Much better work in this field appears to result from a small number of people working for a long period of time than from a large massive crash effort attempting to generate a complete system within short time''Software for ALIS falls into two main categories: data analysis software (Section 6.2) and software for controlling ALIS. In the early phase of work on ALIS, initial ideas for how to develop the software for controlling ALIS were discussed by Nilsson and Nyström [1992]; Johansson [1991]; Steen et al. [1990] and Nilsson [1994]. While the main ideas for the control centre software have been in use until now, this is not the case for the station software.
ALIS is a complex and highly configurable instrument compared to traditional ground-based auroral imaging equipment. Someone once remarked that ``comparing ALIS with a traditional all-sky camera is like comparing EISCAT with an ionosonde''. ALIS is a campaign-orientated instrument and therefore needs more attention prior to measuring compared to traditional imagers. This gives large amount of flexibility for performing new types of scientific studies, but also introduces new problems. First of all, ALIS is a new instrument, with new and partly unexplored possibilities, and to learn those possibilities takes long time. While absolute measurements from photometers have been analysed for almost half a century, there is no common established practice to deal with such data expanded to two dimensions (not to mention three dimensions). Traditionally images have belonged to the realm of morphological measurements (some exceptions exist). Another concern is the tremendous increase in the amount of data produced, and how to process and analyse the data in a way that enables and improves the possibilities for physical interpretation. The point of this discussion is simply that it is difficult to develop a control system for something that is to a large extent new and unknown. Other difficulties arise from technical limitations, as well as from the rapid technical development, causing some components to become obsolete before they are finished and making it difficult to put priorities in the right order. (In the beginning of the project various proprietary programming environments had to be procured and learned, often at a high cost. Some years later, more standardised, reliable and free software emerged, partly making the earlier efforts superfluos. If that had been known in advance, those resources could have been spent on other tasks.) The development of a control system for ALIS was therefore carried out by successive approximation and ``survival of the fittest'' ideas. The following sections will provide a brief overview of a vast subject that is best described by running ALIS and studying the source code.