Toward the adaptation of component‐based architectures by model transformation: behind smart user interfaces

Abstract

SUMMARYGraphical user interfaces are not always developed for remaining static. There are GUIs with the need of implementing some variability mechanisms. Component-based GUIs are an ideal target for incorporating this kind of operations, because they can adapt their functionality at run-time when their structure is updated by adding or removing components or by modifying the relationships between them. Mashup user interfaces are a good example of this type of GUI, and they allow to combine services through the assembly of graphical components. We intend to adapt component-based user interfaces for obtaining smart user interfaces. With this goal, our proposal attempts to adapt abstract component-based architectures by using model transformation. Our aim is to generate at run-time a dynamic model transformation, because the rules describing their behavior are not pre-set but are selected from a repository depending on the context. The proposal describes an adaptation schema based on model transformation providing a solution to this dynamic transformation. Context information is processed to select at run-time a rule subset from a repository. Selected rules are used to generate, through a higher-order transformation, the dynamic model transformation. This approach has been tested through a case study which applies different repositories to the same architecture and context. Moreover, a web tool has been developed for validation and demonstration of its applicability. The novelty of our proposal arises from the adaptation schema that creates a non pre-set transformation, which enables the dynamic adaptation of component-based architectures.

Exploring Quality-Aware Architectural Transformations at Run-Time: The ENIA Case

“…The metrics presented here are focused on our particular domain of component-based systems, but they can be adapted according to the needs. This step requires three inputs: the set of alternative software architectures (generated by the default model transformation process), a set of metrics to measure QAs and constraints (see Table 1 for QAs and Table 2 for constraints), and the specification of the components to feed the metrics at run-time (stored in the metamodel [8,10]). It is important to emphasize that Tables 1 and 2 show metrics elicited and validated by an expert in the domain architecture of component-based systems, but they may be adapted to the needs of other domains.…”

“…Furthermore, UIs in ENIA are reconfigured at runtime with the aim of adapting their structure to the user interactions, profile preferences, context changes and pro-active system decisions. Since UIs are represented by models, this adaptation process is based on model-to-model (M2M) transformations of component-based architectures (see [10] for further details).…”

“…Previous studies have shown that model transformation is a good approach to adapt the component-based architectures [10]. Existing transformation processes focus on the functionalities of systems, giving less importance to the Quality Attributes (QA).…”

“…Model transformation is a common approach to adapt the component-based architecture of software systems [10]. In this context, Model-Based Engineering (MBE) techniques facilitate the development of software architectures, defining them (including the structure, components' specifications, and run-time interaction) by models.…”

“…Our goal is to modify the transformation schema proposed in [10] for generating more than one alternative and consider quality information to select the best transformation.…”