ConfiLog: An approach to Guide Product Line Configuration Processes based on Configuration Traces

Houssem Chemingui (1, 2) , Ines Gam (1, 2) , Raul Mazo (1, 3, 4) , Camille Salinesi (1) , Henda Ben Ghézala (2)

1 CRI - Centre de Recherche en Informatique de Paris 1
2 RIADI - Laboratoire de recherche en Génie Logiciel, Applications distribuées, Systèmes décisionnels et Imagerie intelligente [Manouba]
3 Universidad EAFIT - Grupo GIDITIC- Medellin, Colombia
4 Lab-STICC_ENSTAB_CID_DECIDE

[Context] Product line engineering is a conception and production paradigm in which the purpose is no longer to develop a single product but to develop a collection of products. They present a strategy that provides to the organizations a more competitive edge by improving productivity and decreasing costs. However, these benefits can be missed if users find difficulties when they are configuring a product from the line. This situation is widespread mainly when they deal with large product lines that contain a plethora of characteristics. [Problem] In these cases, the configuration process becomes monotonous and an error prone task. Furthermore, it is crucial to guide stakeholders by providing assistance during such complex process. Consequently, the configuration process envisages recommending the best configuration alternatives to users until leading them to a satisfying experience. [Contribution] Our proposal aims to enhance future configuration processes and maximize the user satisfaction based on previous executions. Therefore, we mine and analyze previous product line configurations by logging user's actions. User guidance is orchestrated by a goal-question-metric approach. [Results] Our research has shown that guidance solutions can be proposed according to the trace mining. Integrating these solutions throughout the configuration process can assist potential users that they are configuring complex lines. The usability of analyzing traces in order to guide future configurations is demonstrated using an e-shop case study from the software industry.

Mots clés: Product line engineering; configuration process; assistance; recommending; user's actions; goal-question-metric; trace mining

Domaines: Informatique [cs]; Génie logiciel [cs.SE]

Houssem Chemingui, Ines Gam, Raul Mazo, Camille Salinesi, Henda Ben Ghézala. ConfiLog: An approach to Guide Product Line Configuration Processes based on Configuration Traces. 2019. ⟨hal-02295969⟩