Research / Methodology
v1.0 · 2026
How a result
becomes a
claim.
The institutional and technical layers that govern Grunuss research outputs — process, methods, validation discipline, and the whitepaper template that every release must satisfy.
§ 01 / Premise
GS-2026 / SECT_01
Methodology is the verification mechanism.
Methodology at Grunuss governs how an investigation moves from hypothesis through internal review and methodological audit to external publication. The process is invariant; the technical work within it evolves with the science. Together, the institutional and technical layers define how the institution earns the right to make claims.
The same template that governs a public whitepaper governs the technical memo that precedes it. Drift between internal informality and external rigor is treated as a structural failure, not a presentation choice. This page documents the standards every output must satisfy before it leaves the institution.
§ 02 / Layers
GS-2026 / SECT_02
One continuous standard.
Methodology has an institutional layer and a technical layer. The institutional layer is invariant — it governs how a result moves from investigation to public release regardless of the technical work that produced it. The technical layer is the computational stack that produces the work, and evolves with the science. Each layer has its own discipline; together they form the verification mechanism.
- L.01Institutional layerStages of investigation, peer review, audit, and release. The invariant process. Documented in § 03 below.
- L.02Technical layerNamed computational methods — electronic-structure modelling, multiscale physics coupling, predictive materials analysis, validation against measurement. Documented in § 04 below.
§ 03 / Institutional process
GS-2026 / SECT_03
Institutional process — from investigation to release.
Research outputs follow a defined progression. Each stage has a documented exit criterion. No output reaches the external layer without passing every prior stage.
- IP.01Internal investigationHypothesis is formalised. Assumptions are declared in writing before computation begins. The investigation is logged with parameter sets, model references, and the conditions under which a result would be considered invalid.
- IP.02Structured peer reviewResults are reviewed by domain colleagues not involved in the investigation. Review covers method, assumption integrity, statistical handling, and replicability. The review record is preserved alongside the result.
- IP.03Methodological auditBefore external release, the work is audited against the institutional whitepaper template (§ 06 below). Sections that cannot be filled honestly are not filled. A whitepaper with an empty Limitations section is not released.
- IP.04External publicationRelease follows the whitepaper template documented in § 06. The published form is the form that was audited — promotional reframings are not permitted.
§ 04 / Technical methods
GS-2026 / SECT_04
Technical methods.
The technical stack reflects the three architectural pillars. Specific methods are named below; specific parameter sets accompany each released result.
- TM.01Electronic structure modellingDensity Functional Theory (DFT) and post-DFT methods for band structure, defect formation energies, and electronic transport properties. Used to define the lower bound of physical possibility for engineered material systems.
- TM.02Multiscale physics couplingLinkage between atomistic simulation, mesoscale microstructural modelling, and continuum finite-element analysis. Coupling boundaries are explicit; transferred quantities are documented.
- TM.03Predictive materials analysisSurrogate models trained on validated simulation outputs to expand the searchable design space. Surrogate-derived results are flagged as such and never confused with first-principles outputs.
- TM.04Validation against measurementSimulation outputs are compared against experimental or literature-validated reference data wherever such data exists. When it does not, the result is published with that absence stated.
§ 05 / Validation discipline
GS-2026 / SECT_05
Validation discipline.
Per the PAaaS interface doctrine, simulation outputs are not free-form. They must satisfy a fixed set of reporting requirements before they enter any document — internal or external. The rules apply to charts, tables, and prose equally.
- VD.01Parameter listingEvery output declares the parameter set under which it was produced. No output is released without its parameter trace.
- VD.02Model version referenceThe exact software and model version that produced the output is recorded. Results across versions are never silently compared.
- VD.03Timestamp and convergence stateWhen the computation was run and whether it converged are both recorded. Non-converged results are labelled, not discarded silently.
- VD.04Explicit success or failure criteriaWhat would count as a successful run is declared in advance. Post-hoc reframing of failure as success is treated as a methodological error.
- VD.05Uncertainty bounds where applicableWhere a result has a quantifiable uncertainty, the bound is published. Where it does not, the absence is declared.
- VD.06Data visualisation integrityNo truncated axes without explicit notation. No 3D charts. No gradient fills, glow, or neon. Charts prioritise interpretability over visual differentiation.
§ 06 / Whitepaper template
GS-2026 / SECT_06
Whitepaper template.
Every Grunuss whitepaper conforms to the template documented below. The template is published here for external reviewers, collaborators, and prospective contributors. Conformance is not optional. Sections that cannot be filled honestly are not filled, and the whitepaper is not released until they can be.
WT.01
Title Page
Title, authors, institutional affiliation, publication date, version reference, and document identifier. The title states what was done — never what is hoped.
WT.02
Executive Summary or Abstract
A neutral summary of the question, the method, the validated result, and the limitations. Length is bounded. Marketing language is rejected at the audit stage.
WT.03
Institutional Context
How the work relates to the institutional mission. Establishes that the question being asked is within the mandate of the institution, not a tangential exploration.
WT.04
Methodology
The procedure followed, expressed at a level that a domain reviewer could reproduce. Software versions, model references, and parameter sets are named explicitly.
WT.05
Assumptions
A dedicated section. The conditions under which the result holds are enumerated. Assumptions that were considered and rejected are also recorded.
WT.06
Simulation or Modelling Framework
The computational or analytical framework used, with version identifiers and convergence criteria. Where multiple methods were considered, the selection rationale is documented.
WT.07
Results
The validated output of the investigation. Uncertainty bounds are explicit. Charts follow the data visualisation doctrine — no truncated axes, no gradient fills, no 3D, no glow.
WT.08
Validation
Comparison against measurement, literature, or independent simulation. Where validation reference data does not exist, the absence is declared rather than concealed.
WT.09
Limitations
A dedicated section. The boundaries of the result are stated. A whitepaper without an honest Limitations section is not released. This rule has no exceptions.
WT.10
Future Work
Adjacent questions opened by the result, and the conditions under which they would become tractable. Speculation about commercial outcomes is not in scope.
WT.11
References
Full citation of every external work referenced. Internal references are listed by document identifier to support institutional traceability across releases.
Source — §7.6.1 Whitepaper Template Requirements and §7.3.2 PAaaS Interface Doctrine, Grunuss Institutional Philosophy Document. Conformance is audited prior to external release.
Methodology is how the institution earns the right to publish.
The institutional layer is invariant. The technical layer evolves with the science. Together they govern every release that bears the Grunuss name.