Lithography systems for sub-0.25micrometers designs will probably require masks with very high data content, small feature sizes, and extreme accuracies. Moreover, the mask substrates themselves will probably be exotic, compared with today's quartz blanks. We examine the requirements set forth in wafer lithography technology roadmaps and the published characteristics for proposed lithography tools to extrapolate to the mask pattern generation requirements. We then examine the implications for the maskmaking tools in the year 2000. These requirements lead to a discussion of the Etec Excalibur e-beam mask writer program. Finally, we comment on the prospects of an e-beam direct write technology applicable to the year 2000 production requirements and discuss the potential of some architectures proposed in the literature. We intend to show why radical innovation will be required above and beyond what has been disclosed to date.