Disciplines / Research questions

软件工程

10 个研究主题 · 28 个问题

测试覆盖率与现场故障率之间的弱相关之谜4 个问题
  1. Quantitative relationship between structural test coverage and field software reliability

  2. Does Unit-Tested Code Crash Beyond Eclipse? Generalizing and Explaining the Absent Association Between Unit Testing and Field Failures

  3. Which test-suite properties predict post-release defects when confounders are controlled? Generalizing the Apache Commons evidence

  4. Is the association between unit-test coverage and single-statement bugs robust, causal, and generalizable?

大型语言模型生成代码的长期可维护性3 个问题
  1. The long-term maintenance burden of autonomous-agent-generated code: trajectory, composition, and interpretation relative to human-authored code

  2. Validating LLM-derived cross-entropy as a signal of code maintainability

  3. Validated and Enforceable Maintainability and Readability Measurement for LLM-Generated Code

代码审查是否降低缺陷还是仅延迟合并1 个问题
  1. Empirical evaluation of process levers for code review speed: selective commit-then-review, reviewer incentives, and dedicated review time

技术债务自报标记与真实维护成本的对应关系3 个问题
  1. Quantifying the lifecycle and maintenance impact of deep-learning-specific self-admitted technical debt on larger, more diverse corpora

  2. Objective Measurement of Technical-Debt Interest: Validating Perceived Estimates of TD-Induced Wasted Development Time

  3. Do Self-Admitted Technical Debt Types Differ in Their Impact on Method-Level Software Maintenance?

开源项目公交因子与实际维护韧性不一致4 个问题
  1. Threshold calibration and data-driven validation of CST and RIG bus-factor estimation

  2. Why some open-source projects survive losing their core development team: causal mechanisms and measurement robustness of truck-factor developer detachment

  3. A comprehension-grounded instrument for measuring theory retention in software systems at whole-system and whole-team scale

  4. Open validation and extension agenda for the multimodal bus factor estimator of Jabrayilzade et al. (ICSE-SEIP 2022)

缺陷预测模型跨项目迁移性能骤降2 个问题
  1. Source-project selection and exploitation for cross-project defect prediction under heterogeneous metrics and downstream aggregation

  2. Label-free detection of tendency-reversed metric matches in heterogeneous defect prediction

微服务拆分后开发敏捷性与运维复杂度的反转3 个问题
  1. Causal outcome evaluation of microservices migration decision-points: quantifying how migration choices affect migration outcomes

  2. How do the motivations, benefits, and issues of microservices adoption evolve over time? An unresolved empirical question from Taibi, Lenarduzzi and Pahl (2017)

  3. Testing methodologies and database-management strategies for monolith-to-microservice migrations: closing the empirical gap identified by practitioner research

需求变更波动如何通过架构依赖放大4 个问题
  1. Generalizing the requirements–architecture dependency taxonomy and change-impact metrics beyond a single case: cross-domain and cross-view validation

  2. Causal relationships between requirements volatility and software architecture outcomes: from single-case correlation to validated causal, cross-context knowledge

  3. Validating the change propagation probability metric for assessing design quality of software architectures on larger case studies

  4. Empirical validation and operationalization of the nine-type requirement dependency model for change propagation analysis

依赖更新机器人是否减少还是增加供应链风险2 个问题
  1. Do automated dependency-update bots on net reduce or increase software supply-chain risk?

  2. Causal attribution, long-term persistence, and heterogeneity of GitHub's security interventions (security alerts, Dependabot security updates, and CodeQL code scanning)

自动程序修复中的过拟合补丁问题2 个问题
  1. Breaking in-loop test–fix coupling in iteratively hardened co-generated bug reproduction tests and fixes

  2. Determining Intended Behavior and Validating Patch Correctness for Under-Specified Issues in SWE-bench-style Benchmarks

返回开放问题