Low-code application platforms enable citizen developers to autonomously build complete applications, such as web applications or mobile applications. Some of these platforms also offer support for reuse to facilitate the development of similar applications. The offered mechanisms are usually elementary, they allow module reuse or building a new application from a template. However, they are insufficient to achieve the industrial level reuse necessary for software product lines (SPL). In fact, these platforms were conceived to help build standalone applications, not software families and even fewer software product lines. In this paper, we argue that the major limitation is that these platforms seldom provide access to their metamodel, the access to applications’ models and code is also limited and, therefore, makes it harder to analyze commonality and variability and construct models based on it. An approach is proposed to surpass these limitations: firstly, a metamodel of the applications built with the platform is obtained, and then, based on the metamodel, a domain-specific language (DSL) that can express the models of the applications, including variability, is constructed. With this DSL, users can combine and reuse models from different applications to explore and build similar applications. The solution is illustrated with an industrial case study. A discussion of the results is presented as well as its limitations and related work. The authors hope that this work provides inspiration and some ideas that the community can explore to facilitate the adoption and implementation of SPLs in the context, and supported by, low-code platforms.

Model driven approaches are shifting software development from a code based activity to a model based activity. Models can be refined and transformed from requirements into code specific to a platform. Although several model transformations can occur, they usually take place at a single development stage. In the case of software product lines, and particularly of software factories, the modeling of a system can occur at several stages, for instance, at the software house, at the systems integrator and at the final customer site. Basically, this requires that the model used at a particular stage can be refined at the next stage. In this paper, we explore the issues related to such an approach and we propose model transformation patterns that can be generically applied to models so that they can be used in multi staged modeling approaches. We show how to realize the approach with the Eclipse Modeling Framework and present an insurance case study.

Features have been widely used by the product line community to model variability. They represent the common and variable characteristics of the members of a product line. They are very well suited for the configuration of product line members. Outside the product line community, use cases are also widely used to model the functionality of systems at a similar level of abstraction but from a user perspective. Significant work has been done by several authors regarding the possible relationship between these two perspectives of a system. Nonetheless, this has been done in an informal way. In this paper we explore the relationships between these two perspectives and describe a possible approach to automate the transformation from UML use case to feature models.

Software product lines and related approaches, like software factories, are starting to capture the attention of the industry practitioners. Nevertheless, their adoption outside the research community and big companies is still very restricted. We believe that model-driven approaches, like OMG's MDA, with proper tool support, can bring the advantages of product lines to a broader audience. For this to become a reality, model-driven methods should integrate requirements models into the software development process. In this paper, we discuss the semantics of use case relationships and their formalization using activity diagrams to support variability specification. Particularly, we propose an extension to the LtextendGt relationship that supports the adoption of UML 2.0 use case diagrams into model-driven methods. Our proposal results from our work with 4SRS (4 step rule set), a model-driven method in which use cases are the central model for requirements specification and model transformation