Enhancing database querying skills by choosing a more appropriate interface
Jozef Hvorecký, Martin Drlík, Michal Munk · 2010
To benefit from data stored in growing numbers of database systems, the number of users capable of finding relevant information should also grow. They should be capable of posing requests, formulate relevant queries, and interpret the obtained results. The user interface should allow them to concentrate on the core of the problem without disturbances caused by details of the query language's syntax. We show the results of our study of the relationship between the interface offered by query languages and their users' ability to express their requests in a correct and comfortable manner. Our aim was to analyze the state-of-art in introductory database courses from this point of view, to design and to test an educational methodology appropriate for nonprofessionals. It should help them to achieve an advanced expertise in query design and development. All DBMS's contain SQL but many of them also have a version of Query-By-Example (QBE) language. Their developers likely believe that for non-professionals this sort of interface is more appropriate for creating queries than SQL. Our review of literature showed that the topic was hardly studied in depth. To do so, we designed and developed an educational methodology which addresses the frontier field between design of query languages, HCI, users' mental and conceptual models and performed an educational experiment that tested it. Many of our presumed advantages of QBE have been proved correct. There are statistically significant differences in time, accuracy and users' satisfaction related to query creation. We also found out statistically significant differences in combinations of variables (time, accuracy, and satisfaction) and the level of task difficulty. The result may have a positive influence on nonprofessionals' exploitation of databases. It may lead to designing and developing a combined methodology that would benefit from positives of both: QBE and SQL.