Uma abordagem orientada a modelos para geração de código para sistemas de informação baseados na Web construídos com frameworks

• Autor(a) principal: Almeida, Nilber Vittorazzi de
• Data de Publicação: 2019
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da Universidade Federal do Espírito Santo (riUfes)
• Texto Completo: http://repositorio.ufes.br/handle/10/13730
• topic: Ciência da Computação; Engenharia Web; Desenvolvimento orientado a modelo; Geração de código; FrameWeb; Web Engineering; Model-Driven development; Code generation; subject.br-rjbn

In the field of Web Engineering, there are several methods proposed for the development of Web-based Information Systems (WISs), FrameWeb is a method that aims to develop WISs that use certain types of frameworks in their architecture, proposing the use of models that incorporate concepts of such frameworks at design-time. Such concepts are represented in the language constructs, whose (abstract) syntax is defined by a metamodel, allowing for the method to be extended. This work proposes a code generation tool from FrameWeb models based on an extension of the language’s metamodel. The FrameWeb Generator, target artifact of this work, is able to transform a previously created model based on the FrameWeb metamodel — using the CASE tool FrameWeb Editor — in a structurally complete software project to be used as basis in the development of a Web project. At the end of the process it is possible to observe the advantages of using FrameWeb, since not only the classes and their basic structures are generated, but also the visual part of the project elaborated in the FrameWeb editor, being one of the great features proposed by the method: the representation of a page and its components directly in the model. The FrameWeb Generator was evaluated through laboratory experiments in which projects previously developed by students of a Web Development course had their FrameWeb models built in the editor and the generated code was compared with the original so that the code generation coverage could be measured. Such experiments showed that a high percentage of the code could be generated automatically, freeing developers from the most tedious part of encoding.