Use of Model Driven Engineering in Building Generic FCA/RCA Tools

Within maintenance software methodologies that analyze existing applications, Relational Concept Analysis (RCA) is an efficient approach to build abstractions in any language, using the existing relations between different software artifacts. Nowadays, there are several RCA-based tools, where a critical aspect is the lack of genericity in the FCA mapping to translate input (and output) data to (and from) formal/relational contexts. Most of the tools provide specific translators that can be used only with the analyzed application domain and their code needs to be changed when the framework of analysis evolves. Using Model-Driven Engineering, we propose a generic encoding/decoding process, which performs this translation with only the configuration of an encoder/decoder tool. This approach eases the integration of a FCA/RCA process in a tool and facilitates its usage on a wide range of input data formats.