It has become a common recommendation to computer interface designers that they should “Know the User” (e.g., Rubinstein and Hersh, 1984). This panel discussion will examine the issues that arise from this advice.

Ambiguity in decision tables reduces or eliminates their effectiveness as a control structure in programming languages or as a representation of production rules in an expert system. Rather than expand the table to remove the ambiguity, three weighting schemes for the condition stub are proposed. Rule weights explicitly assign a rule precedence by assigning a weight to each rule. Condition weights implicitly define precedence by weighting each condition. Entry-weighted tables assign precedence by assigning each entry a weight. In all cases, ambiguity is eliminated by the assignment or calculation of unique rule weights within each ambiguous set. The use of weighted decision tables, in addition to improving the programming language interface, provides improved documentation for complex situations.

Many techniques are employed in numbering lines for text editing. The simplest approach uses a single integer that changes each time a line is added or deleted (1). BASIC, in addition to other systems, uses fixed multi-digit numbers bound to each line. This approach has problems: only a fixed number of lines can be inserted between two consecutive lines of text, otherwise the original text must be renumbered. This negates one advantage of line numbers: the ability to compare different versions of the same document.