The plastic surgery hypothesis

By Earl T. Barr, Yuriy Brun, Premkumar T. Devanbu, Mark Harman, Federica Sarro. 2014

249 citations recorded by the source corpus.

Subfields: Information Systems, Software, Cognitive Science & Psychology, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Information Systems

In Software

In Cognitive Science & Psychology

In Signal Processing

Citations in the corpus

A Survey of Machine Learning for Big Code and Naturalness (2017, background): “The “naturalness hypothesis” implies that source code and its properties are predictable and recent research has exploited this fact [19, 61, 81, 126, 168].”
An Empirical Study on Learning Bug-Fixing Patches in the Wild via Neural Machine Translation (2018, methodology): “[7] examined a history of 15,723 commits to determine the extent to which the commits can be reconstructed from existing code.”
Learning to Fix Build Errors with Graph2Diff Neural Networks (2019, background): “Essentially, the copy mechanism is a way to encode into a neural network the insight from prior research on program repair [6, 18] that often code contains the seeds of its own repair.”
History Driven Program Repair (2016, background): “Previous studies have shown that this is adequate for many automated program repair problems [7].”
Context-Aware Patch Generation for Better Automated Program Repair (2018, background): “Applying mutation operators with ingredients at a finer granularity increases the likelihood to include the correct patches in the search space [4, 30].”
Studying the Usage of Text-To-Text Transfer Transformer to Support Code-Related Tasks (2021, result): “Such an assumption has been verified by at least two independent studies [36], [37].”
DeepSim: deep learning code functional similarity (2018, background): “Barr et al. [9] studied patching programs by grafting existing code snippets in the same program and how difficult this grafting process is.”
A Large-Scale Study on Repetitiveness, Containment, and Composability of Routines in Open-Source Projects (2016, background): “Barr et al. [10] reported a high degree of graftability of code changes, providing a foundation for program auto-repairing.”
Automated Program Repair in the Era of Large Pre-trained Language Models (2022, unknown): “One approach is to use the additional information from project-speciﬁc knowledge (i.e., from buggy project itself following the plastic surgery hypothesis [80]).”
An analysis of patch plausibility and correctness for generate-and-validate patch generation systems (2015, methodology): “The uncorrected presence of incomplete or inaccurate information in the scientific literature hampers progress, especially when the results are visible, widely cited, and the foundation for additional research [40, 15, 32, 39, 25, 9, 33, 12, 31].”
Cross-project code clones in GitHub (2018, unknown): “Software is not only repetitive, but also changes in a repetitive way, as found by multiple studies based on different Microsoft projects, Linux, and multiple open source Java projects (Ray et al. 2014; Nguyen et al. 2013; Barr et al. 2014).”
CODIT: Code Editing With Tree-Based Neural Models (2018, methodology): “) used to build the edits are often taken from the existing codebase [4], [5].”
Is the cure worse than the disease? overfitting in automated program repair (2015, background): “GenProg and TrpAutoRepair tackle this challenge using the observation that programs are often repetitive [7, 22] and logic implemented with a bug in one place is likely to be implemented correctly elsewhere in the same program.”
Graph-Based Mining of In-the-Wild, Fine-Grained, Semantic Code Change Patterns (2019, background): “Prior studies have conﬁrmed that software changes are repetitive [2], [25], [28].”
Large Language Models for Software Engineering: Survey and Open Problems (2023, background): “Furthermore, Barr et al. [44] found that 43% of commits to a large repository of Java projects could be reconstituted from existing code.”