Haas explained that the first problem would require researchers to work on real problems, to participate actively in technology transfer and deployment, to work with the relatively small number of technology providers, and to strive for simplicity. The issue of system management spanned many sub-issues, including monitoring of large systems, simplifying operation from an end-user's standpoint, and reducing total cost of ownership. Meaning to reinforce this point, Haas said that most users were not very smart, and that a good goal would be to make computers that mothers could use. Mary Baker objected to this characterization vehemently, noting that some mothers like herself were quite computer-savvy. After acknowledging his faux pas, the speaker went to the third problem, reliability. He observed that the main issues here were system crashes, fault isolation, and interoperability.
Turning to wide-area networking, he noted that this was one of the few remaining problem areas for performance in modern systems. Better OS support, and rethinking many of the classic distributed systems primitives like RPC in a Web world were some of the other issues under this topic. The emerging field of pervasive computing, and the need for a ``lingua franca'' for interoperability were related challenges. Finally, Haas noted that there were many contexts in which a system could operate: high vs. low end; many vs. few vs. single node; and high vs. low security. Ideally a single system should be able to configure itself to all the environments.