Towards a domain-specific automated software development process

Abstract

Software development automation is a promising approach for reducing the cost and complexity of software development. However, existing solutions are inflexible and do not offer a systematic approach that integrates smoothly with the software development lifecycle(SDLC). This thesis proposes a framework for implementing software development automation solutions based on processes, systematic component reuse and structured user input. To achieve this, this approach considers different application domains as different sub-problems to solve in an automated manner that share common tooling. The advantages lie in the proximity to the SDLC and the engineering aspect of software development, while also ensuring vendor lock-in is eliminated and additional cognitive load is kept at a minimum. It is based on domain models reusage, automated decision making for trade-offs regarding architecture and a suite of tools which take care of code generation of simple, repetitive parts such as integrating third-party software and data, and also receive user input with predefined structure based on existing domain knowledge. A tool was developed in order to demonstrate the potential of the approach, specifically in Big Data Systems, demonstrated in a case study based on relevant industry experience.