National science and technology (S&T) systems are often mentioned as a condition for competitiveness of high technology sectors. Therefore, public S&T policies should actively support the development of national S&T systems. In particular in Eastern Europe an active S&T policy is often demanded to support the development of the supposed domestic "high technology potential". This paper shows that this hypothesis is ill-founded in the case of the software sector. With an industrial economic analysis of the software sector it is shown, that a S&T policy is widely not able to fulfil this expectation. The analysis of the different market segments: standard and individual software, shows that the competition is carried out on axes which can widely not be influenced by a S&T policy. The links between software enterprises and the S&T systems are very weak, which is the result of the conditions of software development and the competition axes used in the software industry. Therefore, only few, and very general, starting points remain for an active S&T policy. Main starting points are: the improvement of the education in modern software technology, improvement of patent protecting laws and their enforcement, and introduction of standardisation procedures and quality standards.