01 / Trace every reported finding
Every research brief links to its primary paper by DOI. Source authors, venue, paper date, source-check date and the material consulted are stated. Abstract-based coverage is explicitly identified; it is not represented as a full-text assessment.
02 / Keep evidence types distinct
The evidence profile describes the methods reported in the source. Theory, numerical simulation and experimental measurement are different forms of support. Peer review is a publication status, not proof of correctness or independent replication.
03 / Preserve the comparison
Numerical gains belong to a reference, a specimen and test conditions. Relative improvements are not rewritten as universal performance. Frequency, wavelength, response threshold and propagation medium remain explicit when they matter.
04 / Separate findings from interpretation
The finding section summarizes what the source reports. Boundary notes prevent unwarranted transfer to other systems. “Our engineering reading” is editorial interpretation, not a new measurement or a statement attributed to the authors.
05 / State how this edition was prepared
This launch edition contains seven AI-assisted summaries checked against accessible primary-source material. The newsroom has not independently reproduced these experiments. No external expert review is claimed. None of the briefs in this edition is sponsored.
06 / Identify illustrations and dates
The lattice image is an AI-generated conceptual illustration. It is not experimental evidence or a representation of the featured specimen. Brief publication dates are separate from paper publication dates. This is a selected research edition, not an exhaustive or continuously updated literature feed.
07 / Make corrections visible
Material factual changes should carry a dated correction on the affected brief, identifying what changed. A source retraction or correction should be reflected in the evidence profile. This first edition has no logged corrections.