The mashup component description language

Abstract

Mashups can be seen as the result of software composition applied to the Web. One of the characteristics of mashup development is the heterogeneity of its building components in terms of logical layering (e.g., user interface, application logic, and data), access method (e.g., REST, SOAP), and composition technique (e.g., scraping vs. clipping, synchronous vs. asynchronous interaction, discrete vs. streaming). This poses a challenge towards the design of mashup tools aiming at lowering the barriers of mashup development, as this heterogeneity needs to be abstracted. In this paper, we address this challenge by proposing a new JSONbased domain-specific language for describing heterogeneous mashup components, called the Mashup Component Description Language (MCDL). MCDL lies at the core of a meta-model for mashup component modeling, and can be used for component discovery and classification but also for user-centric mashup development as it decouples the interface of a mashup component from its underlying implementation technologies.

Conceptual Development of Custom, Domain-Specific Mashup Platforms

“…Therefore, the focus on data mashups only is too narrow and the language has already grown very complex (without concrete vendor adoption). Aghaee and Pautasso [2011], on the other hand, focus on the design of a generic mashup component description language and do not elaborate on their composition into mashups. Their model contains technology aspects (e.g., component wrappers), that we think are instead a runtime aspect that should not affect the component description.…”

Integrating open services for building educational environments

“…Mashups can be seen as the result of applying software composition techniques to the development of web applications [1]. Much effort is being spent nowadays to propose models enabling rich integration of the different component functionalities.…”

Natural End-User Development of Web Mashups

“…For example, Listing 1 conforms to the NaturalMash CNL and describes an enhanced music video search mashup that employs Last.fm (http://www.last.fm/api) to first search for a song and then uses the results to accurately search for the corresponding music videos of the song in YouTube (https://developers.google.com/youtube/). The underlying implementation of the CNL accommodates an abstract component model [22] that: (i) gives a unified technology-neutral description of Web APIs, and (ii) models them in an abstract textual form (natural language description). The abstract component model distinguishes two types of functionality provided by Web APIs, namely, Task -a passive atomic operation, and Event -an active source of control.…”