Source code comprehension strategies and metrics to predict comprehension effort in software maintenance and evolution tasks - an empirical study with industry practitioners
Bibliographic record
Abstract
The goal of this research was to assess the consistency of source code comprehension strategies and comprehension effort estimation metrics, such as LOC, across different types of modification tasks in software maintenance and evolution. We conducted an empirical study with software development practitioners using source code from a small paint application written in Java, along with four semantics-preserving modification tasks (refactoring, defect correction) and four semantics-modifying modification tasks (enhancive and modification). Each task has a change specification and corresponding source code patch. The subjects were asked to comprehend the original source code and then judge whether each patch meets the corresponding change specification in the modification task. The subjects recorded the time to comprehend and described the comprehension strategies used and their reason for the patch judgments. The 24 subjects used similar comprehension strategies. The results show that the comprehension strategies and effort estimation metrics are not consistent across different types of modification tasks. The recorded descriptions indicate the subjects scanned through the original source code and the patches when trying to comprehend patches in the semantics-modifying tasks while the subjects only read the source code of the patches in semantics-preserving tasks. An important metric for estimating comprehension efforts of the semantics-modifying tasks is the Code Clone Subtracted from LOC(CCSLOC), while that of semantics-preserving tasks is the number of referred variables.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".