Ismayle S. Santos¹, Magno L. J. Souza², Michelle L. L. Carvalho², Thalisson A. Oliveira ¹, Eduardo S. Almeida², Rossana M. C. Andrade¹

¹ Computer Science Department - Federal University of Ceara (UFC) - Fortaleza, Brazil

Email: {ismaylesantos,thalissonoliveira,rossana}@great.ufc.br

² Computer Science Department - Federal University of Bahia (UFBA) - Salvador, Brazil

Email: {magnosouza, michellecarvalho, esa}@dcc.ufba.br

Context variability modeling is a key task to deal with Dynamically Adaptive Software. Several variability modeling techniques have been proposed to support this task using Dynamic Software Product Lines. Context-aware Feature Model (CFM) has been identified as a suitable technique to deal with context variability. However, it has limited expressiveness to specify real world constraints related to context information. Considering that design faults can lead to failures at runtime, we extended CFM to tackle its limitations as well as to integrate it with a model checking approach to identify common design faults (see Figure 1).

Our proposal for context modelling and errors prevention follows the steps:

• STEP 1: The software engineer should model the DAS by using the eCFM
• STEP 2: S/he should apply the model checking to identify faults on DAS design. If faults are found, then the eCFM should be updated and checked again.
• STEP 3: If there is no faults, the software engineer has an expressive and correct context variability model

We developed a tool called FixTure eCFM to support our approach and the steps proposed were whole or partially automated. More details of FixTure eCFM are available here.