SEPLDec 4, 2017

A Quantitative Study of Java Software Buildability

arXiv:1712.01024v151 citationsHas Code
Originality Synthesis-oriented
AI Analysis

This addresses a practical issue for researchers, students, and practitioners who rely on building third-party software, providing quantitative evidence for what was previously anecdotal.

The study tackled the problem of build failures in third-party Java software by automatically building over 7,200 open source projects, finding that more than 38% of builds ended in failure, with dependency-related errors being the most common.

Researchers, students and practitioners often encounter a situation when the build process of a third-party software system fails. In this paper, we aim to confirm this observation present mainly as anecdotal evidence so far. Using a virtual environment simulating a programmer's one, we try to fully automatically build target archives from the source code of over 7,200 open source Java projects. We found that more than 38% of builds ended in failure. Build log analysis reveals the largest portion of errors are dependency-related. We also conduct an association study of factors affecting build success.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes