Extracting static and dynamic model elements from textual specifications in humanities

Information systems are composed of different dimensions of information: methodological, structural, architectural, etc. These dimensions appear connected in textual specifications created at early stages of information systems conception. These multidimensional textual specifications are particularly relevant in cases of information systems conceived to support processes in narrative-based domains, such as humanities or social sciences. Working on these domains, a hard effort of dimension decoupling and a fully understanding of the domain is required. Thus, software analysts put a considerable effort in the identification of the most relevant concepts of the domain---to create the static structure of the system---and, in a separate manner, in the identification of the activities---to understand how this information is produced and managed by users in real practice---. This paper presents how existing studies aim to bridge this modelling gap between static and dynamic aspects in information systems from a comprehensive perspective. It also proposes a pipeline method to deal with this gap, using a semi-automatic analysis of textual sources. The pipeline focused on the identification of activities and static concepts, producing concept maps as static models and activity logs as dynamic models. The final goal of the proposed pipeline is reducing the domain understanding complexity required to software analysts in humanities information system conception and improving the quality of the models in both static and dynamic dimensions.