Search

This paper posits that a theoretical model of analyst effort in tracing tasks is necessary to assist with study of the analyst. Specifically, it is clear from prior work by numerous research groups that the important factors in such a model are: the amount of time it takes for an analyst to vet a given candidate link and the amount of time it takes an analyst to find a missing link. This paper introduces a theoretical model of analyst effort as well as a simplified model. A number of simulations were undertaken in order to build effort curves to assist in evaluating numerous tracing scenarios, such as determining at what point in time an analyst should switch from vetting candidate links to manually searching for links not in the candidate list. [ABSTRACT FROM AUTHOR]

The challenges of implementing successful and cost-effective traceability have created a compelling research agenda that has addressed a broad range of traceability related issues, ranging from qualitative studies of traceability users in industry to very technical and quantitative studies. Unfortunately, advances are hampered by the significant time and effort needed to establish a traceability research environment and to perform comparative evaluations of new results against existing baselines. In this panel we discuss ongoing efforts by members of the Center of Excellence for Software Traceability (CoEST) to define the Grand Challenges of Traceability, develop benchmarks, and to construct TraceLab, an extensible and scalable visual environment for designing and executing a broad range of traceability experiments. [ABSTRACT FROM AUTHOR]