Recommendation and Weaving of Reusable Mashup Model Patterns for Assisted Development

Abstract

With this article, we give an answer to one of the open problems of mashup development that users may face when operating a model-driven mashup tool, namely the lack of modeling expertise. Although commonly considered simple applications, mashups can also be complex software artifacts depending on the number and types of Web resources (the components) they integrate. Mashup tools have undoubtedly simplified mashup development, yet the problem is still generally nontrivial and requires intimate knowledge of the components provided by the mashup tool, its underlying mashup paradigm, and of how to apply such to the integration of the components. This knowledge is generally neither intuitive nor standardized across different mashup tools and the consequent lack of modeling expertise affects both skilled programmers and end-user programmers alike.In this article, we show how to effectively assist the users of mashup tools with contextual, interactive recommendations of composition knowledge in the form of reusable mashup model patterns. We design and study three different recommendation algorithms and describe a pattern weaving approach for the one-click reuse of composition knowledge. We report on the implementation of three pattern recommender plugins for different mashup tools and demonstrate via user studies that recommending and weaving contextual mashup model patterns significantly reduces development times in all three cases.

Mining and Quality Assessment of Mashup Model Patterns with the Crowd

“…Before their use, patterns were checked by an expert to ensure their meaningfulness and reusability. The extension is called Baya, and our user studies demonstrate that recommending model patterns has the potential to significantly lower development times in model-driven mashup environments [Roy Chowdhury et al 2014].…”

“…Pipes and our pattern recommender Baya [Roy Chowdhury et al 2014]. The components and mashups supported by these tools can be modeled as follows: let C L be a library of components of the form c = name, IP, IF, OP, emb , where name identifies the component (e.g., RSS feed or Filter), IP is the set of input ports for data-flow connectors, IF is the set of input fields for the configuration of the component, OP is the set of output ports, and emb ∈ {yes, no} tells whether the component allows for the embedding of other components or not (e.g., to model loops).…”

Designing service-based applications in the presence of non-functional properties: A mapping study

“…Model [57], [36], [90], [62], [83], [169], [7], [99], [95], [21], [32], [204], [184], [150], [175], [23], [209], [156], [191], [9].…”

“…Model-Driven [47], [130], [176], [217], [16], [4], [129], [18], [195], [21], [63], [87], [51], [156], [9].…”

Crowd-Based Mining of Reusable Process Model Patterns

“…Before their use, patterns were checked by a mashup expert assuring their meaningfulness and reusability (e.g., see Figure 1 for an example of a good pattern). The extension is called Baya [13], and our user studies demonstrate that recommending model patterns has the potential to significantly lower development times in model-driven environments [12].…”