The research

Substrate
Geometry

One method, four programmes. Classical geometric primitives iterated into real engineering materials; static-descriptor QSAR architectures tested where they are known to break; difficult mystical texts rendered under a declared hermeneutic frame; and the structure of the prime gap sequence, probed against pre-committed nulls. Different fields, the same discipline: commit the thresholds before running anything, then publish what the data actually says.

This is the peer research behind the byline that publishes the Lattice network. It is why the network's editorial standard is checkable rather than claimed.

Every paper & preprint ORCID 0009-0005-6869-308X Google Scholar
The programmes
Physics · geometry

Substrate Geometry

Three preprints live on arXiv · two papers under journal review

Taking classical geometric primitives — the oloid, the Gömböc, mono-monostatic bodies — and iterating on them to improve real engineering substrates.

Computational rolling-contact mechanics that classifies shapes by physical invariants rather than by appearance: seven physics oracles, a 25.6× mean-temperature advantage on the TPMS heat-exchanger geometry, and a 58× improvement in contact distribution.

The programme produced the first openly verified Gömböc construction and a 13-member verified mono-monostatic catalog, both released with their computation rather than as a claim.

Paper I was endorsed on arXiv by Georg Nawratil (TU Vienna), after Hellmuth Stachel — who originated the oloid problem — read the work and referred it onward to him. The methodology paper is under review at the SMAI Journal of Computational Mathematics; the TPMS heat-exchanger paper is under review at Elsevier Results in Engineering.

Chemistry · computational toxicology

hERG QSAR

Paper I live on ChemRxiv · Paper II in moderation

The same method pointed at a second substrate: testing static-descriptor QSAR architectures against psychedelic-class compounds, where they are known to be weakest.

Paper I — iboga alkaloids and architecture-specific failure modes — is publicly live on ChemRxiv at DOI 10.26434/chemrxiv.15003259/v1, with its OSF pre-registration public at 10.17605/OSF.IO/UWVX4. The pre-registration is the point: the thresholds were committed before anything was run.

Paper II is the three-family data-landscape contribution across 59 psychedelic-class compounds — tryptamines, phenethylamines and cathinones. Submitted to ChemRxiv 2026-05-24 and in moderation.

Tradition · computational translation

Computational Translation

Three papers · one under review at Aries (Brill)

The same commit-and-audit discipline turned on the problem of rendering difficult mystical texts with a declared hermeneutic frame rather than an invisible one.

Three papers sit on this side: a reception-lens experiment on a passage of Plotinus, the first public-domain English of the canonical Heikhalot Rabbati, and a design case-study on translation scaffolding — the last under review at Aries (Brill).

Where the physics work commits thresholds, the translation work commits a frame: the interpretive stance is declared up front and carried by a two-witness trail, so a reader can disagree with the reading rather than only with the reader. It is the engine behind Hekhal and its /targum bilingual pipeline.

Number theory

Prime-Gap Probe

Manuscript on file · not yet released

A bounded characterization study asking whether the prime gap sequence carries spectral and topological structure beyond what its first moment imposes.

Two methodologically independent observables — a graph-Laplacian spectral dimension benchmarked against Watson 2025, and a persistence-diagram topological signature — each with a single pre-committed gate threshold and three null baselines (Poisson, GUE, uniform).

A null result is as publishable as a positive one, and the gates were written down before the computation ran. Adversarially reviewed across v3–v6 scoping iterations. Target: math.NT.

The method

The same four rules run in every field.

The bet the whole operation runs on is that a careful method does real work in fields that never talk to each other. It is also the reason the commercial network can be held to a standard: the editorial rules on a Lattice property are these rules, applied to a buying decision instead of a rolling body.

Commit the thresholds first
Every study pre-registers what would count as a failure before anything is run. A null result is a publishable result, which is the only condition under which a positive one means anything.
Publish what the data says
Not what would sell. Where sources conflict, no figure is published at all until they are reconciled — the same standard the network applies to its own commercial pages.
Release the computation
Constructions and catalogs ship with the code that produced them, so a reader can disagree with the result rather than only with the author.
One resolvable byline
Every paper and every network article resolves to the same ORCID-anchored identity on the public web. That single @id is what makes the credential checkable instead of asserted.

This is the standard the network is held to.

The Lattice network publishes original, DOI-backed data studies across its properties under this same byline and this same method. If you are evaluating the network as a publisher — or considering a placement on it — the research is the part you can audit without asking anyone for permission.

The Lattice network The open datasets About the operator