From 462b1275dff349a8e9ee341b294c9b2f11d0b6c1 Mon Sep 17 00:00:00 2001 From: Rafael Vuijk Date: Tue, 27 Oct 2026 03:36:27 +0000 Subject: [PATCH] A convention put to the maintainer comes with what the references say AGENTS.md already asks for every convention to be checked against a second and third source, Wolfram MathWorld and Wikipedia among them, before it is adopted. It did not say the same of a question: on #2131 the notation choices a geometry book raises were put to the maintainer as bare alternatives, and he asked for Wikipedia and MathWorld to be consulted first. Read, they settled most of them: (a, b) as the pair, dot and cross as a . b and a x b, the plane's pseudo-scalar product as MathWorld's perp dot product, and the Kronecker delta as an Iverson bracket. The section now says that a question about a notation, a name, a convention or a definition carries what those sources say, quoted and linked, and a recommendation. Co-Authored-By: Claude Opus 5.5 (1M context) Claude-Session: https://claude.ai/code/session_012sonx8iAspMiwRwokT1Ura --- AGENTS.md | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/AGENTS.md b/AGENTS.md index 2ecc3d6f2..4ae73fdf4 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -433,6 +433,12 @@ list above, Wolfram MathWorld, Wikipedia's notation tables — and say which was Terminology the sources themselves call non-standard (MathWorld on *natural number*, *whole number*, *counting number*) is a reason to pick the unambiguous spelling, not the familiar one. +**That holds for a question as much as for a decision.** A notation, a name, a convention or a +definition put to the maintainer comes with what those sources say about it, quoted and linked, +and a recommendation that follows from them. Bare alternatives are not a question, only reading +left undone. On [#2131](https://github.com/asc-community/AngouriMath/issues/2131) the notation +choices were first asked bare; with Wikipedia and MathWorld read, most of them answered themselves. + The output of the reading is a document with those lists — objects, notation, operations, methods, what is wrong, test cases — kept beside the goal it came from, and the docket of work is derived from it. A summary written before the reading is finished is the thing this section