back to table · edit · history · where entries came from · files · download
3747 bytes, as of the version from 2026-11-20 00:19 (current). Recorded here, not run.
# Argument selection for T768
Instruction source: ../20261120T000421Z-build-skills/manifest.json, archived matching local/live complete bundle b5c3c7985b7c9e178b4899d083a9e3bc4284989dff76f05e5b0e52835d47eed2; core b79b7fbc9fa3ce63f2d642cccdac12b0d214757ab01db66c03a2d84264a58e5f. This is builder justification, not independent acceptance.
The sole free parameter k is the actual nonnegative real coefficient of xyz in x^4+y^4+z^4+1+kxyz. It is continuous (type R), not a discrete index. The natural family includes 0<k<4, although no such value is selected in this run. Ordinary logarithmic measure uses natural logs and Haar probability, with integrable zeros and no regularization. Rogers (0704.2438v4), Eq. (2.6), fixes f4(u)=4 B(u^(1/4)); Samart (1205.4803v2), Prop. 1.3(iii), establishes the exterior formula including u=256. Root-of-unity changes do not change the measure, so the positive fourth-root convention is consistent.
Exhaustive, disjoint mixed groups of machine identities:
* Intrinsic: `0`, `4`. The first is the Fermat degeneration, carried by the degree-four torus map to m(1+x+y+z). Four is the contact boundary of the torus triangle bound and the singular quartic fibre. Rogers Cor. 2.6 and Samart Prop. 1.3(iii) supply the endpoint control.
* Arithmetic structure (published CM comparisons): `648^(1/4)`, `2304^(1/4)`, `12`, `28`. These are precisely Samart Theorem 1.4, Eqs. (1.9)–(1.12). Their fourth powers are 648, 2304, 20736, 614656. The first root is uniquely positive in (5,6), the second in (6,7). The 12 row belongs to this group even though it also lies in the integer slice.
* Exploratory reference sample: `5`, `6`, `7`, `8`, `9`, `10`, `11`, `13`, `14`, `15`, `16`. Together with the already assigned 4 and 12 these exhaust the integer band 4<=k<=16 of primitive integer-coefficient quartics. Each polynomial is primitive because four fixed coefficients equal 1. These are plausible recognition targets among integral members of this specific arithmetic pencil, not a claim of CM or an application for each integer. The bound is four times the intrinsic torus-contact scale; sixteen is explicitly a noncanonical practical cutoff independent of decimal notation. It is neither a probability grid nor a multidimensional coefficient mesh. The checked Rogers and Samart passages support the family and convention, not use of every exploratory argument.
Reference policy: purpose = recognition of simple integral quartic coefficients beyond the exceptional CM cases; bounds = every integer coefficient from 4 through 16, after intrinsic/CM overlaps are assigned above; coverage rationale = one fixed polynomial pencil, spanning its full first exterior coefficient range to four times contact while retaining the intrinsic degeneration and all four source CM cases; noncanonical_acknowledged = true; notation_independent = true; sources_checked = true.
Breadth precedes depth: both the Fermat and contact cases are retained, ordinary integral members are not pruned to famous-only cases, and all source-selected CM arguments are included with exact identities. Interior and complex values require further ordinary-measure analysis and are future extensions of this table, not new quantities. This selection does not create an exponential, a derivative-period or a determinant table.
Default enumeration, stored entries, Programs and complete-note give exactly 17 distinct identities: 0, all integers 4..16, the two nonintegral positive fourth roots, and 28. The parameter domain is k>=0; complete=no. No dependencies or input files control the generator's selection. A fresh independent critic must evaluate this choice and bind the live document and generate.py with agents/argument_review.py fingerprint.