Business intelligence and analytics software serve several purposes in ubiquitous computing environments because they are incorporated into almost all devices used for communication, carrying out transactions, and providing feedback. Unfortunately, designing context-aware software that can adapt to their immediate environment is extremely challenging in dynamic business environments, thus presenting several issues to business and analytic system efficiency and user control. Although program developers can easily develop application software that specifies how user interfaces should behave, it is nearly impossible to designate the contexts of these adaptations or create UIs that interact with their environment and satisfy individuals’ needs. Thus, Ghaibi et al. presented a technique known as the Model Driving Engineering Approach, which allows developers to design UIs that can respond to their environments and allow the end-users full control over adaptations during run time.
According to the authors, the technique is based on a graphical tool and conceptual framework that simplifies the work of designers when creating adaptive end-user interfaces, editing rules, modelizing context situations, and managing the whole process. Ghaibi et al. (2018) explain that rule-based approaches in UI adaptations are attractive due to their simplicity. Unfortunately, they cannot adapt to new system data and require professional control. Moreover, the rules applied to the UI framework remain fixed as they cannot be altered by machine learning. In other words, the only solution to designing a fully adaptive system using rule-based approaches is to describe all possible situations during the design stages, which is relatively impossible. Thus, the authors suggest equipping the system to identify and recognize new contexts and rules and build on its present set-up to produce desirable end-user results. Thus, they concluded that machine learning approaches fit perfectly with the problems as they can convey user interface adaptations.
The approach adopted by the researchers is straightforward and relatively simple to follow through. Ghaibi et al. (2018) adopted a Model-Driven-Development approach to create a Final-User Interface, designed to respond to specific situations. However, they integrated the adaptation specifications into the Camelon Reference Framework to ensure consistency when making changes at any level of the representation or regenerating the interface. As a result, the solution is comprised of a toolkit that offers graphical tools to aid UI generation from task levels to the final level, context situation parameter management, adaptation management, and launching. Ghaibi et al. (2018) suggest that these tools ensure full control over the UI adaptation process, thus allowing developers and designers to configure the interface and its components according to their user’s preferences or software capabilities. However, the main issues addressed by the Model-Driven-Development approach are the complexity of UI systems since it aims to reduce efforts used in developing UI systems. Nonetheless, Ghaibi et al. (2018) emphasize adopting a modeling approach with varying abstraction levels, achievable using model-model or model-code transformations. However, the authors exploit several benefits of this approach, especially at the design and generation stages.
An advantage of the proposed solution is the definition of models to describe various context components, including the UIs environment, platform, and user, as well as the recommended rules of adaptation. The description of context components is a notable advantage since editing rules was not supported or addressed by a majority of existing and previous approaches. Moreover, there are situations when developers and designers define a single contextual situation using various rules, thus introducing conflicts within the system. In addition, conflicts within the UI can distort the consistency of requirements and disrupt or even limit the functioning of other adaptations (Ghaibi et al., 2018). Thus, defining the models also played a crucial role in simplifying the process and avoiding issues in the final presentation.
Reference
Ghaibi, N., Dâassi, O., & Ayed, L. J. B. (2018). User interface adaptation based on a business rules management system and machine learning. Communications of the IBIMA.