Comment ne plus créer de bugs ?

posted on 28/11/2025

Bug : défaut dans un programme informatique entraînant un dysfonctionnement.

Catégories

# Bug Type (label) Definition / Description Frequency (among the 1 280 bug reports) Time-to-fix / Fixing-process Observations
1 Configuration issues Bugs due to mis-configuration: e.g. wrong build setup, incorrect external-library versions or dependencies, wrong paths in config/manifest files, incorrect configuration parameters. (ResearchGate) One of the more frequent types. (azaidman.github.io) Has its own characteristic fixing timeline (distinct from other bug types) — time before response, assignment, change, and full fix differ compared to other types. (azaidman.github.io)
2 Network issues Bugs caused by network or inter-component communication problems: e.g. timeouts, protocol misuses, broken connections, network failures or data transfer problems. (ResearchGate) Less frequent than core categories (functional, GUI, configuration) but still present. (azaidman.github.io) Time-to-fix distribution differs from other types — the fix process shows a distinct profile in how quickly the team responds and resolves. (azaidman.github.io)
3 Database-related issues Bugs where interaction with a database fails: e.g. wrong queries, failed DB connections, issues retrieving or storing data, data corruption or inconsistency due to DB operations. (ResearchGate) Among the less frequent categories in their sample. (azaidman.github.io) Fixing timeline is distinct — from report to response/assignment/change may differ compared to functional or UI bugs. (azaidman.github.io)
4 GUI-related issues Bugs in the user interface: visual/layout problems, incorrect rendering or styling, user-input/UI element misbehavior, graphical or presentation bugs (UI glitches, misaligned controls, wrong UI behavior). (ResearchGate) One of the more common bug types among the data set. (azaidman.github.io) The fix process for UI bugs shows characteristic timings; likely moderate compared to functional and security bugs. (azaidman.github.io)
5 Performance issues Bugs that affect performance: e.g. resource leaks (memory, CPU), unresponsive behavior, slow operations, inefficient code paths, high resource consumption or performance bottlenecks. (ResearchGate) Relatively less frequent compared to central categories but present. (azaidman.github.io) Time-to-fix shows a unique distribution — some performance bugs may linger, require more investigation, exhibit variability in fix duration. (azaidman.github.io)
6 Permission / Deprecation issues (perm-/depr-) Bugs due to the use of deprecated APIs, permission problems, removed APIs or features, or misuse of permissions or outdated interfaces/constructs. (ResearchGate) Among the less frequent types. (azaidman.github.io) Their fix process timing differs; may be sometimes quicker or slower depending on how clear the deprecation or permission issue is. (azaidman.github.io)
7 Security issues (or “security-related issues”) Bugs that pose security vulnerabilities or weaknesses: e.g. misuses of security-sensitive APIs, improper handling of permissions or authentication, security-related flaws. (ResearchGate) These are relatively rare compared to functional/UI/config bugs. (azaidman.github.io) Among the fastest to get an initial response: the study notes that security-related bugs have the smallest “time before response” (TBR), often mean and median equal to 2 (very quick). (azaidman.github.io)
8 Program-anomaly (Functional) issues Core logic bugs: flaws in application logic — wrong return values, exceptions, crashes, incorrect behavior not tied to UI / config / performance / DB / network — basically functional errors in application logic. (ResearchGate) By far the most frequent bug type in the dataset (i.e. functional/program-anomaly bugs dominate). (azaidman.github.io) Developers also tend to respond quickly to functional bugs (short delay before first reaction), likely because they directly impact functionality and reliability. (azaidman.github.io)
9 Test-code–related issues Bugs located in test code rather than production code: e.g. failing tests, flaky or unreliable tests (e.g. tests that pass/fail unpredictably), broken mocks, incorrect test logic or incorrect test setup/execution. (ResearchGate) Less frequent than main categories but part of the 9-type taxonomy. (azaidman.github.io) Test-related bugs also get fairly prompt (relative) attention: despite potentially lower severity, developers react reasonably fast. (azaidman.github.io)