The commercial offer of massively parallel systems has now reached a wide variety, with well engineered products offered by European and American manufacturers. Furthermore, both the American government and the Commission of the European Communities have clearly identified the important role that High Performance Computing may play in science and industry in the near future. To be successful, HPC needs to improve greatly the quality attributes of its software. This paper investigates the status of the software environments for massively parallel systems. In fact, whilst the hardware technology improvements in the past few years have been very rapid, the technology has suffered because of a substantial misalignment between the potential user's expectations and the actual usability of software for massively parallel computer systems. Actions and serious industrialisation efforts are needed to bring the system software and the basic programming environments to a level of real usability, because it is time for these systems to get rid of the ancestral equation: massive parallel systems = number crunching. Many potential markets could be addressed by massively parallel systems if they were able to offer to the users the same ease of use, quality, completeness, and portability of conventional sequential systems.