Bringing Together Configuration Research: Towards a Common Ground
summary
Shows a concrete failure mode: running MegaHit with defaults yields contig counts 284, 285, 285, 289 because kmax is silently reset to 119, documentation lists kmax=141, kalist hides a dependency, and kmax has odd-number and size constraints. Demonstrates Apache “options bleed” caused by a misspelled HTTP method and missing module enabling. Proposes a common model and platform ACCORD and a doxygen-based tooling patch that extracts defaults, constraints and BusyBox doxygen tracing to give machine-readable config metadata.
AI-written summary · about
chapters (12)
- 0:00Start
- 6:01Configuration Research Everywhere
- 9:24Common Solutions
- 10:45Dimensions of Configurability
- 11:21Example Configuration Bugs
- 13:56ACM CCS
- 19:12A First Step
- 19:32ACCORD: A Community for Configurability Open Research and Development
- 20:15Ideal World
- 21:13Systems Community
- 21:58Scientific Community
- 22:24PL Community
related talks
25mFinding Unmet Dependencies in Kconfig with the Kismet Static Analyzer
15mStatic Detection of Silent Misconfigurations with Deep Interaction Analysis
01h 14mBuilding and Organising a Multi-Platform Development Code Base
52mBringing Existing Open-Source Code into MISRA Compliance
30mA Series of Unfortunate Bugs
43mQuit Simplifying!
14mThe state of Morello software and projects
01h 01mLittle things and big things: lessons for building large systems.