From 439ba58843679ae93bdd148b7ccf2fd661c4ee32 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Alexander=20K=C3=B6lnberger?= <159939812+ProfRandom92@users.noreply.github.com> Date: Wed, 20 May 2026 12:48:29 -0700 Subject: [PATCH 1/2] docs: normalize API surface project naming --- docs/API_SURFACE.md | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) diff --git a/docs/API_SURFACE.md b/docs/API_SURFACE.md index 18fd249..2ce842a 100644 --- a/docs/API_SURFACE.md +++ b/docs/API_SURFACE.md @@ -2,7 +2,7 @@ ## Purpose -This document describes the stable Comptextv7 API, dashboard, export, and report +This document describes the stable CompText V7 API, dashboard, export, and report surfaces that future benchmark/regression summaries can reference. It also records integration expectations for sanitized outputs from `ProfRandom92/Comptext-Daimler-Experiment-` without introducing runtime coupling. @@ -32,7 +32,7 @@ available. This guardrail is intentionally lightweight: it validates helper scripts, available local checks, and contract examples while benchmark/regression evidence remains a sanitized report handoff from `ProfRandom92/Comptext-Daimler-Experiment-`, not a runtime dependency of -Comptextv7. +CompText V7. Branch discipline remains part of the API contract process: create a feature branch from `main` when available, open a PR, request review, and never push @@ -77,7 +77,7 @@ summaries include: ## Export/report endpoints `/export.json` and `/export.csv` are the primary report handoff endpoints inside -Comptextv7. They should remain deterministic enough for review, small enough for +CompText V7. They should remain deterministic enough for review, small enough for PR artifacts, and explicit about any schema changes. `docs/reports/dashboard-health-summary.json` is the dashboard-facing static @@ -94,14 +94,14 @@ Future report endpoints such as `/reports/benchmark-summary`, `/reports/regression-summary`, or `/reports/sanitization-summary` should not be added until a schema, security review, and issue scope approve them. If added, they should consume sanitized summaries only and should not execute experiment -repository workloads from the Comptextv7 runtime. +repository workloads from the CompText V7 runtime. ## Payload and report contract expectations Accepted report summaries should be: - Synthetic in documentation examples. -- Sanitized before being copied into Comptextv7. +- Sanitized before being copied into CompText V7. - Small enough to review in a pull request. - Text-based: Markdown, JSON, or CSV. - Explicit about `source_repo`, `target_repo`, `report_type`, status, timestamp @@ -115,7 +115,7 @@ Machine-readable contract schemas now live under `contracts/` and are written as lightweight JSON Schema-like documents that future agents and CI can inspect without adding runtime dependencies: -- `contracts/api-dashboard.schema.json` describes Comptextv7 API routes, +- `contracts/api-dashboard.schema.json` describes CompText V7 API routes, dashboard views, export formats, sanitized report integration points, and security notes. - `contracts/benchmark-summary.schema.json` describes synthetic benchmark @@ -183,9 +183,9 @@ of reimplementing release-readiness logic. ## Compatibility with benchmark/regression reports -Benchmark summaries should map to Comptextv7 review surfaces this way: +Benchmark summaries should map to CompText V7 review surfaces this way: -| Report type | Expected Comptextv7 use | +| Report type | Expected CompText V7 use | | --- | --- | | `benchmark_summary` | Compare p50/p95/p99, RPS, error rate, and payload size for dashboard/API routes. | | `regression_summary` | Decide whether a PR should merge, require remediation, or split into smaller changes. | @@ -198,14 +198,14 @@ or high/critical forensic findings. ## Dashboard/API boundaries -Comptextv7 may: +CompText V7 may: - Display sanitized benchmark/regression status. - Export local dashboard evidence as JSON or CSV. - Document future report contracts. - Add small schema-version fields in future PRs. -Comptextv7 should not yet: +CompText V7 should not yet: - Import code from `ProfRandom92/Comptext-Daimler-Experiment-`. - Run experiment repository workloads as part of normal dashboard/API startup. From 25b3733ddc083f617412d59aa1d460c9feb8a9d6 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Alexander=20K=C3=B6lnberger?= <159939812+ProfRandom92@users.noreply.github.com> Date: Wed, 20 May 2026 12:49:02 -0700 Subject: [PATCH 2/2] docs: normalize research positioning project naming --- docs/research_positioning.md | 30 +++++++++++++++--------------- 1 file changed, 15 insertions(+), 15 deletions(-) diff --git a/docs/research_positioning.md b/docs/research_positioning.md index 93044cd..dd8e6b5 100644 --- a/docs/research_positioning.md +++ b/docs/research_positioning.md @@ -2,9 +2,9 @@ CompText V7 is the deterministic replay-integrity layer for compressed operational agent traces. It asks whether compact or reconstructed operational state can preserve the evidence, constraints, blockers, dependencies, recovery paths, capability boundaries, and tool-order signals needed to replay a safe operational trajectory after compression. The project is complementary to learned context-compression research, RAG evaluation, vector-memory systems, serving-layer cache optimization, and durable workflow infrastructure, but it does not replace those systems or claim solved AI memory. -## What CompTextv7 measures +## What CompText V7 measures -CompTextv7 measures fixture-bound replay survivability with deterministic artifacts. Current metrics and labels are intended to show whether explicitly encoded operational fields survive replay pressure, not whether a model answer is useful or semantically complete. +CompText V7 measures fixture-bound replay survivability with deterministic artifacts. Current metrics and labels are intended to show whether explicitly encoded operational fields survive replay pressure, not whether a model answer is useful or semantically complete. Measured signals include: @@ -20,15 +20,15 @@ Measured signals include: ## Replay-survivability evaluator brief -CompTextv7 evaluates replay survivability of compact operational state: whether fixture-authored operational fields can be compacted, reconstructed, replayed, and audited without relying on an LLM judge. The current prototype measures field survival, evidence survival, operational drift, and deterministic failure labels against checked-in fixtures. Its claims are therefore fixture-bound and prototype-scoped: it can show what the current validators detect under replay/compression pressure, not whether a deployed agent or memory product will succeed in the wild. +CompText V7 evaluates replay survivability of compact operational state: whether fixture-authored operational fields can be compacted, reconstructed, replayed, and audited without relying on an LLM judge. The current prototype measures field survival, evidence survival, operational drift, and deterministic failure labels against checked-in fixtures. Its claims are therefore fixture-bound and prototype-scoped: it can show what the current validators detect under replay/compression pressure, not whether a deployed agent or memory product will succeed in the wild. -Adjacent benchmark ecosystems include long-term memory benchmarks, RAG evaluation, long-horizon agent evaluation, software-agent/task benchmarks, and context-compression evaluation. Those ecosystems often evaluate task success, retrieval, answer quality, memory recall, or downstream performance. CompTextv7 is complementary: it evaluates whether compact operational state remains replayable and auditable, and it identifies which blockers, constraints, evidence, dependencies, recovery paths, or tool-order signals fail under compression/replay pressure. +Adjacent benchmark ecosystems include long-term memory benchmarks, RAG evaluation, long-horizon agent evaluation, software-agent/task benchmarks, and context-compression evaluation. Those ecosystems often evaluate task success, retrieval, answer quality, memory recall, or downstream performance. CompText V7 is complementary: it evaluates whether compact operational state remains replayable and auditable, and it identifies which blockers, constraints, evidence, dependencies, recovery paths, or tool-order signals fail under compression/replay pressure. -Why this matters: fluent summaries can lose blockers, constraints, evidence, dependencies, or recovery paths while still reading well. CompTextv7 treats that as deterministic replay degradation, not subjective text quality. The review path is the current trust chain: fixtures, generators, committed artifacts, Markdown summaries, README/doc values, artifact drift validation, and CI checks. See [Iterative Replay Degradation](iterative_replay_degradation.md), [Benchmark Explanation](BENCHMARK_EXPLANATION.md), the committed [iterative replay degradation summary](../artifacts/iterative_replay_degradation_results.summary.md), and [`scripts/validate_replay_artifact_drift.py`](../scripts/validate_replay_artifact_drift.py). +Why this matters: fluent summaries can lose blockers, constraints, evidence, dependencies, or recovery paths while still reading well. CompText V7 treats that as deterministic replay degradation, not subjective text quality. The review path is the current trust chain: fixtures, generators, committed artifacts, Markdown summaries, README/doc values, artifact drift validation, and CI checks. See [Iterative Replay Degradation](iterative_replay_degradation.md), [Benchmark Explanation](BENCHMARK_EXPLANATION.md), the committed [iterative replay degradation summary](../artifacts/iterative_replay_degradation_results.summary.md), and [`scripts/validate_replay_artifact_drift.py`](../scripts/validate_replay_artifact_drift.py). -## What CompTextv7 does not measure +## What CompText V7 does not measure -CompTextv7 does not measure general intelligence, answer quality, production readiness, or universal memory. It intentionally avoids: +CompText V7 does not measure general intelligence, answer quality, production readiness, or universal memory. It intentionally avoids: - LLM judges or subjective scoring; - embeddings, vector databases, graph stores, and external APIs; @@ -50,23 +50,23 @@ The core contribution is a small deterministic review layer for operational repl ## Operational state vs raw chat history -CompTextv7 focuses on operational state, not raw chat-history retention. Rather than preserving every dialogue turn, it extracts, compacts, reconstructs, and verifies the fields that fixtures declare operationally relevant: tasks, constraints, blockers, evidence, dependencies, tool order, recovery actions, and continuation requirements. +CompText V7 focuses on operational state, not raw chat-history retention. Rather than preserving every dialogue turn, it extracts, compacts, reconstructs, and verifies the fields that fixtures declare operationally relevant: tasks, constraints, blockers, evidence, dependencies, tool order, recovery actions, and continuation requirements. This framing is intentionally narrower than semantic memory. A replay can pass only for the fields represented in the fixture and checked by the deterministic validator. ## How deterministic replay validation differs from adjacent categories -| Category | What that category usually evaluates or provides | CompTextv7 boundary | +| Category | What that category usually evaluates or provides | CompText V7 boundary | | --- | --- | --- | -| RAG evaluation | Retrieval quality, answer grounding, citation coverage, or generated-answer quality. | CompTextv7 does not retrieve documents or judge generated answers. It checks whether fixture-defined operational state survives compact/replay cycles. | -| Vector memory | Embedding-based recall and similarity search over stored memories. | CompTextv7 does not use embeddings or vector databases. It compares explicit fixture IDs, fields, attachments, and normalized values. | -| KV-cache compression | Serving-layer efficiency for model attention/cache reuse. | CompTextv7 does not optimize model internals or inference caches. It emits reviewable replay artifacts and field-survival metrics. | -| Workflow orchestration | Durable execution, retries, scheduling, state machines, and tool execution. | CompTextv7 does not run autonomous workflows. It validates whether replayed operational state still contains fixture-defined continuation requirements. | -| Learned context compression | Model-learned summaries or compressed prompts optimized for downstream performance. | CompTextv7 does not train or evaluate a learned compressor. It measures deterministic replay preservation under controlled fixtures. | +| RAG evaluation | Retrieval quality, answer grounding, citation coverage, or generated-answer quality. | CompText V7 does not retrieve documents or judge generated answers. It checks whether fixture-defined operational state survives compact/replay cycles. | +| Vector memory | Embedding-based recall and similarity search over stored memories. | CompText V7 does not use embeddings or vector databases. It compares explicit fixture IDs, fields, attachments, and normalized values. | +| KV-cache compression | Serving-layer efficiency for model attention/cache reuse. | CompText V7 does not optimize model internals or inference caches. It emits reviewable replay artifacts and field-survival metrics. | +| Workflow orchestration | Durable execution, retries, scheduling, state machines, and tool execution. | CompText V7 does not run autonomous workflows. It validates whether replayed operational state still contains fixture-defined continuation requirements. | +| Learned context compression | Model-learned summaries or compressed prompts optimized for downstream performance. | CompText V7 does not train or evaluate a learned compressor. It measures deterministic replay preservation under controlled fixtures. | ## Artifact-backed JSON and CI checks -CompTextv7 uses artifact-backed JSON and deterministic Markdown summaries so reviewers can inspect the exact replay evidence for a commit. CI artifacts are evidence records for tested fixtures; they are not universal guarantees. +CompText V7 uses artifact-backed JSON and deterministic Markdown summaries so reviewers can inspect the exact replay evidence for a commit. CI artifacts are evidence records for tested fixtures; they are not universal guarantees. ## Fixture-bound baseline interpretation