docs: add v3 kernel-excellence roadmap anchor#16
Merged
Conversation
The v3 track (Smart Compaction, Cache Aligning, True Background Tasks, Self-Correction) differs from v1/v2 in one structural way: the four items are pairwise coupled, so a decision in one constrains the others. Unlike the independent v1/v2 milestones where "commit message == spec" sufficed, the coupling needs cross-cutting invariants locked BEFORE implementation, or they get violated three PRs later and force rework. docs/v3-kernel-roadmap.md is the anchor. It locks three things: 1. The four capabilities, which existing primitive each builds on, their coupling edges (compaction breaks cache; verification is a background task; everything must stay offline-testable), and the dependency-respecting sequence (NOT 1-2-3-4: cache telemetry first so compaction's effect is measurable, then compaction, then the self-correction loop, background control last). 2. Three cross-cutting invariants that hold for the whole track: compaction must be cache-stable (append, never rewrite a cached prefix); every mechanism must stay offline-deterministic-testable via FakeModel / mock executor; injecting results into context must happen after the cache breakpoint. 3. The already-decided §2 entry milestone: rolling prefix breakpoint, M3.1a (message-prefix cache_control) + M3.1b (hit-rate telemetry), with M3.1c (main-loop fork-trace attribution) deferred to when §1 compaction actually needs the "why did this turn miss" alarm. Pre-spec rigor is tiered by blast radius: §2 is small/reversible -> build directly; §1/§3/§4 get a written design section here first. Docs-only; 187 tests still green. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The v3 track (Smart Compaction, Cache Aligning, True Background Tasks, Self-Correction) differs from v1/v2 in one structural way: the four items are pairwise coupled, so a decision in one constrains the others. Unlike the independent v1/v2 milestones where "commit message == spec" sufficed, the coupling needs cross-cutting invariants locked BEFORE implementation, or they get violated three PRs later and force rework.
docs/v3-kernel-roadmap.md is the anchor. It locks three things:
The four capabilities, which existing primitive each builds on, their coupling edges (compaction breaks cache; verification is a background task; everything must stay offline-testable), and the dependency-respecting sequence (NOT 1-2-3-4: cache telemetry first so compaction's effect is measurable, then compaction, then the self-correction loop, background control last).
Three cross-cutting invariants that hold for the whole track: compaction must be cache-stable (append, never rewrite a cached prefix); every mechanism must stay offline-deterministic-testable via FakeModel / mock executor; injecting results into context must happen after the cache breakpoint.
The already-decided §2 entry milestone: rolling prefix breakpoint, M3.1a (message-prefix cache_control) + M3.1b (hit-rate telemetry), with M3.1c (main-loop fork-trace attribution) deferred to when §1 compaction actually needs the "why did this turn miss" alarm.
Pre-spec rigor is tiered by blast radius: §2 is small/reversible ->
build directly; §1/§3/§4 get a written design section here first.
Docs-only; 187 tests still green.