Integração de fontes de dados convencionais e não convencionas através de uma abordagem WOPA (Write Once, Persist Anywhere)

Abstract

Most computer programs require some mechanism for storing and retrieving data, but these mechanisms do not use to have direct integration with programming languages. This integration requires too much programming effort, which led to the development of several works dealing with this problem. The most common approaches are based in object-relational mapping, language-integrated query and other solutions that are not flexible enough. Thus, the resulting code is linked to a specific set of resources, requiring the full rebuilding of a big part of it. This paper presents a solution for abstracting data sources, allowing them to be mapped through libraries and settings outside the source code in an interoperable manner. As result, we attempted to achieve the increase of code reuse, the decoupling of data sources, the reducing of rework and the increasing of productivity, without limiting the expression capability of the programmer.