01 / Information
Equal records, different responses.
Shear and aspect-ratio pairs establish analytic transmission-error floors when an allowed predictor receives the same reduced record.
Research / Open working paper
Geometry is information.
What does a compact description of an opening need to retain to predict the light passing through it?
Separate the model
from its approximation.
The study considers ideal straight parallelogram recesses with absorbing walls under spatially uniform angular illumination. Within this model, a reduced geometry record can omit information needed to predict total transmission.
That is different from a numerical method retaining the geometry but approximating its response poorly. The two problems call for different remedies.
Change the shape while keeping its area fixed. This is a geometry illustration, not a transmission calculator.
Working revision 7 · 27 October 2026
Geometry information and approximation cost in recessed-aperture transport
Mostafa Nasr · Independent researcher, Giza, Egypt
DOI: 10.5281/zenodo.23216693
Working paper · Not peer reviewed
The fixed revision and its supporting code, data, protocols and verification records are available in the public Zenodo archive.
Read the paper and archiveOriginal project code: MIT. Manuscript, data, figures and documentation: CC BY 4.0. See the archive for the applicable license files.
What the work contributes
A bounded comparison.
A next experiment.
02 / Computation
Methods under a declared criterion.
A saved computational matrix compares information ablations, integration, interpolation and harmonic approximations. A separate clipped-domain adaptation helps distinguish different uses of harmonics.
03 / Experiment planning
A comparison worth building toward.
The aspect-ratio pair provides a starting point for a physical test. Geometry and measurement uncertainties must be propagated before interpreting a physical separation.
Evidence has
boundaries.
Analytic statements apply within the model. Reported numerical checks come from executed programs and saved results. Empirical refinement indicators are not certified error enclosures.
No physical measurement or achieved instrument accuracy is reported. The manuscript discloses substantial AI assistance and its evaluation limits. Independent technical comments and later bench results can inform subsequent work.
Follow the physical build