Automação de feedback para melhorar a nomeação de identificadores de alunos.

Abstract

Program readability is a fundamental tenet of software quality. Thus, providing timely feedback on identifier naming can help to improve novice programmer training in programming education. Indeed, several studies in software engineering literature indicate that a source code that contains well-chosen identifier names is easier to understand and less error-prone in maintenance tasks. However, due to the numerous students are learning to program nowadays in programming courses, providing manual feedback on identifier name quality become prohibitive due to the highest workload imposed on instructors. During this master research, we proposed an innovative method to generate and provide automated feedback on identifier name appropriateness assessment in contrast to the checking of naming conventions. Our challenge was to provide to students timely and elaborated feedback to help them to choose more appropriate names to denote their code identifiers. We implemented and evaluated our proposal in an introductory programming course. Firstly, we investigated and verified that it is possible to automatically find names considered to be inappropriate code identifiers, with the intent of using this information to generate automated feedback. In sequence, we investigated and witnessed that we can generate and give automated feedback to students so that we can help them to choose better quality code identifier names. The main contribution of this research is that we can provide automated feedback to students in order to encourage them to improve names chosen by them to denote their code identifiers. So, we can help students to improve identifier naming and, in consequence, their program readability, since their early coding experiences.