Relational concept explored

a consensus among the panelists that high-quality software is the key to successful mini systems (especially on-line ones), audience questions were focused on how the inexperienced or potential mini user can get the first-class system definition and detailed specifications needed to produce a successful system. Consultants can fill this rote well and should be able to offer objectivity in evaluating and implementing hardware and software offerings, Vic Poor said. But he pointed out that qualified consultants of integrity are so few and far between that exploding demand for software applications will require most users to depend on reliable hardware manufacturers offering total systems capability. Pan-elists from systems and software firms, however, felt that independently supplied software has filled a vendor support void, can satisfy users particular needs more precisely and can employ the hardware best suited to a particular job. In summary, minis are already providing a surprisingly powerful range of low-cost solutions for major business problems. The seminar on transferability of application programs and data bases at NCC 76 included presentations by Robert Marion of the Defense Communications Agency, Elmer Bartels of Honeywell and James Fry of the University of Michigan. The session highlighted the users' needs for translation aids, currently available software provided by the manufacturers and results and directions of related research activities. The paper presented by Fry and coauthored by Edward Birss was selected as the best technical paper of NCC 76. "Relational data base" is a term which is slowly beginning to be used in business data processing circles. Most practitioners still feel that there is something vaguely ivory-towerish, esoteric and probably impractical about the whole concept. After all, EDP people had to rapidly adjust to the linear, to hierarchical, to network transitions; now we are throwing yet another data base model at them. That was the theme advanced for a seminar on relational data bases at NCC 76. The discussion, however, suggested that relations are not esoteric, they can be practical, real people are solving real problems in the area. Three papers were presented: one on the University of Ontario's virtual memory system (RAP), one on General Motor's combination of DB, graphic and statistical package (REGIS) and one on IBM's fill-in-the-blanks (Query by Example). Though at first glance, it seemed that the three topics were quite divergent, they did have a unifying message, that relational data bases are real, they are usable and there …