Retrospective: Data Mining Static Code Attributes to Learn Defect Predictors

Tim Menzies · 2025

Read Retrospective: Data Mining Static Code Attributes to Learn Defect Predictors by Tim Menzies (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-15662.txt.

Description

Tim Menzies published this retrospective note in 2025, looking back at one of software engineering's most cited papers. The original study mined simple static measures of program code, such as size and complexity counts, to learn predictors of where defects would appear. It released its data and scripts, drawn from NASA projects, so anyone could reproduce the baseline result.

The note traces what happened next. Hungry for shared data, the research community embraced the paper, and at its 2016 peak it was the field's most cited paper per month. By 2018, a fifth of leading journal papers used artifacts from this line of work. Menzies reflects on which lessons from that runaway success deserve to be kept, and which should be forgotten. The note also carries a broader claim: industry can get research it wants by publishing baselines with data.

Catalog record

CREATORTim Menzies
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE23 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-15662

Browse the full library · About the Tower of Babel