Skip to content

Latest commit

 

History

History
60 lines (43 loc) · 6.05 KB

File metadata and controls

60 lines (43 loc) · 6.05 KB

Decision Log

ID Status Decision Reason
D-001 Accepted Use SujitSaiY2007/embedded-runtime-resilience as the canonical project repository. Provides durable project continuity across chats and sessions.
D-002 Accepted Keep the final topic unfrozen until prior-art research is completed. Prevents confirmation bias and unsupported novelty claims.
D-003 Accepted Restrict the target implementation to software-based embedded systems. Core project constraint.
D-004 Accepted Treat publication and patent potential as evaluation goals, not guarantees. Novelty and patentability must be established through evidence and professional assessment.
D-005 Accepted Maintain explicit handoff/state documents in the repository. Enables continuation from future chats without reconstructing the project from conversation history.
D-006 Accepted Insert a formal research-space decomposition step before candidate-specific prior-art selection. Prevents anchoring on the current lead and makes alternative opportunity classes comparable.
D-007 Accepted Advance six opportunity classes to systematic prior-art investigation, with OC-02, OC-01, and OC-10 as the initial investigation order. They combine real embedded problems, feasible MCU implementation, measurable baselines, and plausible research depth; this ranking is not a novelty claim.
D-008 Accepted Treat the runtime-contract/recovery candidate as provisional and potentially too broad. Phase 0 found substantial existing work in runtime verification, watchdogs, resource contracts, and local recovery.
D-009 Accepted Freeze the development topic as a lightweight context-aware peripheral recovery policy for resource-constrained event-driven MCU firmware. Deep prior-art attack found the broad peripheral-recovery space crowded but identified a narrower, experimentally testable software-only policy hypothesis worth developing.
D-010 Accepted Treat context-aware recovery, selective peripheral restart, graceful degradation, and recovery state machines as established concepts rather than novelty claims. Academic, industrial, and patent evidence contradicts broad novelty claims.
D-011 Accepted Define the provisional contribution around local fault context/history, bounded recovery-action selection, and explicit preservation of unrelated event-driven service. This is the narrowest defensible hypothesis identified in the current investigation and supports measurable comparison against fixed retry/reset baselines.
D-012 Accepted Integrate a strict zero-heap constraint into the frozen development topic. Static allocation is appropriate for deterministic resource-constrained MCU firmware; zero-heap itself is not claimed as novel.
D-013 Accepted Integrate event quarantine as a core recovery mechanism, while using formal invariants rather than claiming full formal verification. Quarantining only the fault-associated event while preserving independent valid queued events directly operationalizes the service-preservation objective and creates measurable correctness properties.
D-014 Accepted Treat MPU support as a target-platform containment mechanism, not as the invention itself. MPU technology is established; it can strengthen isolation experiments without becoming a false novelty claim.
D-015 Accepted Require physical MCU hardware for final physical validation, while permitting laptop/PC simulation and development before hardware is obtained. MPU behavior, peripheral-driver behavior, interrupt timing, and real hardware fault modes cannot be fully established by host-only simulation.
D-016 Accepted Transition to Phase 1 — System Design and Experimental Preparation. Topic is frozen; remaining work is to design the smallest defensible system and reproducible evaluation before implementation.
D-017 Accepted Make MPU-capable MCU/board selection an immediate Phase 1 prerequisite. Physical validation requires real MCU behavior; board choice affects architecture, peripherals, fault injection, and tooling.
D-018 Accepted Treat event model, fault model, recovery state machine, formal invariants, baselines, and experimental protocol as design gates before large-scale implementation. Prevents framework-first development and keeps the project experimentally falsifiable.

Frozen development topic

Design and Implementation of a Lightweight Zero-Heap Context-Aware Peripheral Recovery Policy with Event Quarantine for Resource-Constrained MPU-Enabled Event-Driven MCU Firmware

Short working title: Zero-Heap Context-Aware Peripheral Recovery with Event Quarantine

Rejected/Deferred decisions

R-001 — Premature topic freeze

Status: Superseded by D-009 after the requested deep prior-art attack.

The topic is now frozen for development, but not certified novel or patentable.

R-002 — Immediate large-scale implementation

Status: Deferred.

Implementation begins after the exact architecture, fault model, baselines, and experimental methodology are defined.

R-003 — Treat the broad runtime-contract/recovery framework as already novel

Status: Rejected.

The Phase 0 evidence scan indicates substantial prior art around runtime verification, watchdog/recovery, resource contracts, and fine-grained recovery.

R-004 — Treat generic peripheral recovery as the invention

Status: Rejected.

Phoenix, Karma, driver-recovery research, vendor recovery frameworks, and patent activity show that generic peripheral recovery and selective recovery are established.

R-005 — Freeze a broad claim of context-aware selective recovery

Status: Rejected.

Context-aware health management and recent selective peripheral-recovery patents make the broad concept insufficiently differentiated.

R-006 — Treat zero-heap, MPU use, or event quarantine individually as novel

Status: Rejected.

Each is an established technology/concept. The research contribution must come from the specific integrated mechanism and its demonstrated technical trade-offs.