Does feature scattering follow power-law distributions?
Bibliographic record
Abstract
Feature scattering is long said to be an undesirable characteristic in source code. Since scattered features introduce extensions across the code base, their maintenance requires analyzing and changing different locations in code, possibly causing ripple effects. Despite this fact, scattering often occurs in practice, either due to limitations in existing programming languages (e.g., imposition of a dominant decomposition) or time-pressure issues. In the latter case, scattering provides a simple way to support new capabilities, avoiding the upfront investment of creating modules and interfaces (when possible). Hence, we argue that scattering is not necessarily bad, provided it is kept within certain limits, or thresholds. Extracting thresholds, however, is not a trivial task. For instance, research shows that some source-code-metric distributions are heavy-tailed, usually following power-law models. In the face of heavy-tailed distributions, reporting metrics in terms of averages and standard deviations is unreliable, although commonly done so. Thus, prior to extracting reliable thresholds for feature scattering, one must understand the shape of feature-scattering distribution. In this direction, we analyze the scattering degree of five C-pre-processor-based software families and verify whether their empirical cumulative feature-scattering distributions follow power laws. Our results show that feature scattering in the studied subject systems have characteristics of heavy-tailed distributions, with a good-fit with power laws. Hence, we raise awareness that feature scattering thresholds based on central measures may not be reliable in practice.
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 | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| 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".