UML and Multiplicities Under a Microscope

• In Event: SET System Design

Abstract

Using an eye tracker and computerized experiment, the effect of several variations in descriptions of business rules on multiplicity validation performance is examined. We find that cognitive performance during multiplicity validation is affected by a number of factors. Based on the chosen performance metric, semantic models account for an increase in effort of 107-112% compared to syntactic models. In semantic model interpretation tasks, we find that semantic variation increases task difficulty on the order of 44-45%. The interpretation of a minimum multiplicity is more difficult than the interpretation of a maximum multiplicity resulting in approximately 1 second and 4 eye fixations of task interference. We also find that it is easier for individuals to validate a syntactic model, but in the presence of semantics invalidating a model takes less effort. Our findings have the potential to inform theory and practice on how model design decisions impact usability.

Authors

• Walter R. Boot, Florida State University

• Cheryl L Dunn, Grand Valley State University

• Bachman P Fulmer, cal state university - fullerton

• Gregory J Gerard, Florida State University

• Severin Grabski, Michigan State University