From 6ec0e964c9ef15b8055dbfcf380dc784e793913b Mon Sep 17 00:00:00 2001 From: "Joshua (D) Drake" <136637981+ChronicallyJD@users.noreply.github.com> Date: Thu, 6 Aug 2026 09:26:39 -0600 Subject: [PATCH 1/2] docs: record what the ClickBench licence actually forbids (#421) Owner determination, 2026-08-06, after reading CC BY-NC-SA: publishing our own test results is permitted. We measure to compare, not to sell. An earlier version of this commit removed the ClickBench numbers from docs/benchmarks.md on a reading that no numbers could be published. That reading was wrong and the removal is reverted. Nothing about those measurements was in doubt at any point; only what we were entitled to do with them. What the term forbids, now written down rather than left to inference: - selling copies of the material, or of a remix of it - placing it on a site monetised by ads where revenue is the primary goal - using it in promotional materials, corporate brochures or paid advertisements - charging a fee to access, download or view a derivative work built on it The third is the one worth guarding, because it is the one a well-meaning reader walks into. A benchmark table in documentation exists so a user can decide whether this engine suits their workload. The same table in a sales deck is the prohibited use, and having produced the numbers ourselves does not change that, because the definition they were produced with is the licensed material. So the caution sits beside the table in docs/benchmarks.md, where someone about to lift it will see it, and the full list sits in PROVENANCE.md beside the determination and its reasoning. No legal opinion was sought, and it is recorded so it can be revisited rather than re-derived. --- PROVENANCE.md | 37 ++++++++++++++++++++++++++++++++----- docs/benchmarks.md | 11 +++++++++++ 2 files changed, 43 insertions(+), 5 deletions(-) diff --git a/PROVENANCE.md b/PROVENANCE.md index 3667acb..1c8338a 100644 --- a/PROVENANCE.md +++ b/PROVENANCE.md @@ -551,11 +551,38 @@ oracle, so none changes query results. this repository. The runner fetches `postgresql/create.sql` and `postgresql/queries.sql` from upstream at run time, into the benchmark data directory, and feeds them to `psql` unmodified. The comparison oracle is the - heap arm of the same run, not any upstream expected output. **Proposed - 2026-08-05: keep the run-time fetch rather than take a durable in-tree copy. - Owner decision pending, and the NonCommercial term's effect on publishing - benchmark numbers is a separate question that a run-time fetch does not - address.** The measured numbers published in + heap arm of the same run, not any upstream expected output. + + **Both open questions were decided by the owner on 2026-08-06.** + + *Run-time fetch:* keep it, and re-fetch on every run rather than caching, so a + run measures the current upstream definition rather than whatever was current + the first time the harness ran on that machine. + + *NonCommercial:* the owner read the licence and determined that **publishing our + own test results is permitted**. We measure to compare, not to sell. What the + term forbids, and what this project must therefore not do with ClickBench or + anything derived from it: + + - selling copies of the material, or of a remix of it + - placing the material on a website or platform monetised by ads where the + primary goal is revenue generation + - using the material in promotional materials, corporate brochures, or paid + advertisements for a business + - charging a fee to access, download, or view a derivative work built on it + + The third is the one to watch, because it is the one a well-meaning reader walks + into. The numbers in `docs/benchmarks.md` are documentation: they exist so a user + can decide whether this engine suits their workload, and so we can see our own + regressions. **Lifting them into a brochure, a sales deck, a conference booth + panel or a paid campaign is the prohibited use**, and the fact that we produced + the numbers ourselves does not change that, because the benchmark definition + they were produced with is the licensed material. + + Recorded as the owner's determination and its reasoning. No legal opinion was + sought. It is written down so it can be revisited rather than re-derived. + + The measured numbers published in `docs/benchmarks.md` are our own, produced on our own hardware. The dataset (`hits.tsv.gz`) is downloaded for local measurement and is not redistributed; its own licensing is unestablished and it must not be added to the tree without diff --git a/docs/benchmarks.md b/docs/benchmarks.md index 8b5c7ea..cd66bf2 100644 --- a/docs/benchmarks.md +++ b/docs/benchmarks.md @@ -739,6 +739,17 @@ project's MIT license does not carry. The harness fetches the definition from upstream at run time and copies nothing. Our comparison oracle is the heap arm of the same run. +The definition is re-fetched on every run, so a result is only comparable to +another taken against the same upstream definition. + +**These numbers are documentation, and the NonCommercial term governs what else +they may be used for.** Publishing our own test results here is permitted: we +measure to compare, and to catch our own regressions. Reusing them in promotional +material, a corporate brochure, a sales deck or a paid advertisement is not, and +producing the numbers ourselves does not change that, because the benchmark +definition they came from is the licensed material. The full list of prohibited +uses is in `PROVENANCE.md` beside the owner's determination of 2026-08-06. + The numbers below are one run on 2026-08-05. The conditions were PostgreSQL 18.4 non-assert, 16 cores, 62 GB of memory, and 11,110,833 rows. That row count is every ninth row of the real 100 million row table. The reported time is the best From ac8b350eeb91e5d229e6c4688d7b3cf7e968575b Mon Sep 17 00:00:00 2001 From: "Joshua D. Drake" Date: Thu, 6 Aug 2026 18:21:50 -0600 Subject: [PATCH 2/2] docs: bring the licence caution inside the STE sentence limit (#421) docs_style failed on the paragraph this PR adds: two sentences of 34 and 37 words against a 25-word limit. Split into short ones, with no change of meaning. The bold run also has to end on a plain period. `used for.**` is not a sentence boundary to the splitter, so the caution and the sentence after it counted as one 34-word sentence rather than two short ones. Co-Authored-By: Claude Opus 5 (1M context) Claude-Session: https://claude.ai/code/session_01L2DvnWDM7g27ubDCQdXhky --- docs/benchmarks.md | 16 +++++++++------- 1 file changed, 9 insertions(+), 7 deletions(-) diff --git a/docs/benchmarks.md b/docs/benchmarks.md index 917eceb..d4c059b 100644 --- a/docs/benchmarks.md +++ b/docs/benchmarks.md @@ -762,13 +762,15 @@ run was on 2026-08-05, and upstream carried the digests above a day later. That is an inference and not a measurement. The next run is the first that will state it. -**These numbers are documentation, and the NonCommercial term governs what else -they may be used for.** Publishing our own test results here is permitted: we -measure to compare, and to catch our own regressions. Reusing them in promotional -material, a corporate brochure, a sales deck or a paid advertisement is not, and -producing the numbers ourselves does not change that, because the benchmark -definition they came from is the licensed material. The full list of prohibited -uses is in `PROVENANCE.md` beside the owner's determination of 2026-08-06. +**These numbers are documentation**, and the NonCommercial term governs their +reuse. Publishing our own test results here is permitted. We measure to compare, +and to catch our own regressions. + +Reuse in promotional material, a corporate brochure, a sales deck or a paid +advertisement is not permitted. Producing the numbers ourselves does not change +that. The benchmark definition they came from is the licensed material. The full +list of prohibited uses is in `PROVENANCE.md`, beside the owner's determination +of 2026-08-06. The numbers below are one run on 2026-08-05. The conditions were PostgreSQL 18.4 non-assert, 16 cores, 62 GB of memory, and 11,110,833 rows. That row count is