Human factors evaluation of two data base query languages: Square and sequel

Abstract

A series of experiments was conducted to evaluate the learnability of two data base query languages, SQUARE and SEQUEL, using university students as subjects. The students were divided into two groups according to whether they had had programming experience. Experiments showed that both populations were able to use either language with reasonable proficiency after 12 to 14 academic hours of instruction (lowest class mean was 55 percent for SQUARE non-programmers; mean score for programmers was 77 percent in both languages). Programmers learned both languages more quickly and more completely than did non-programmers, and the non-programmers showed greater proficiency with SEQUEL than with SQUARE. Test scores showed that both programmers and non-programmers were able to combine the basic language features in ways they had not been explicitly taught. Also, it was shown that, one week after the end of instruction, subjects still retained nearly all their proficiency and were able to use the languages successfully on a closedbook examination. The individual features of the languages varied considerably in learnability. The basic language feature, a simple mapping, was learned in each language with near-perfect accuracy by programmers after two hours and by non-programmers after four hours. However, considerable difficulty was experienced in learning and retaining the more complex "free variable" and "GROUP BY" features, especially by non-programmers. A study of errors made by subjects suggests that a real data base system should be prepared to correct minor syntactic errors and to search for poorly-specified data values by some technique such as stem-matching or a synonym dictionary.