From Products Descriptions to Feature Models

The use of feature models to describe the common and variable features of a family of products is becoming commonplace. Unfortunately, the task of creating a feature model from an existing source is both time-consuming and error-prone, requiring substantial effort from the designer. We developed a semi-automated process as well as tool support for extracting feature models based on products description, tabular data for instance. The tool support is composed of a dedicated language, called PS2FM, for parameterizing through high-level directives (e.g., products/features scoping) the extraction process. We guarantee that the resulting feature model compactly represents the valid combination of features supported by products and has a readable tree hierarchy together with variability information.

Data Corpus