Reproducible
The experiment is not described at a distance. Its inputs, transformations, and results remain available for the reader to rerun.
Reproducible · Auditable · Verifiable
RAVA is an interactive publication format where the manuscript, evidence, and evaluation logic are released as one inspectable object. A claim you can test—not only read.
The format
The format does not replace peer review or rewrite the paper. It creates a faithful interface around the evidence so a reader can encounter the research as a sequence of testable decisions.
The experiment is not described at a distance. Its inputs, transformations, and results remain available for the reader to rerun.
Every interaction resolves back to a source claim, figure, definition, or method. The manuscript stays authoritative.
Readers form a hypothesis before seeing the result, then test whether the evidence supports the published conclusion.
Registry / Issue 001
A ransomware detector reported 100% accuracy. The interactive article asks the reader to remove its shortcuts, change the unit of analysis, and find the honest result.
Machine-learning evaluation · Endpoint telemetry
Publication protocol
A compact editorial method for turning static evidence into a reader-led sequence without loosening its connection to the paper.
Open on the result that deserves scrutiny, not a generic abstract.
Let the reader alter the assumptions, features, and evaluation frame.
Pair every discovery with the manuscript passage that supports it.
Editorial standard
01Every number comes from the source manuscript or its released evidence.
02Every interaction declares what changed and what stayed fixed.
03Every conclusion returns the reader to the authoritative record.
About RAVA
RAVA was created by Kian Esmaeili as a publication interface for research that benefits from direct inspection. The first article uses his own machine-learning evaluation work, so the format is tested against a real manuscript, real evidence, and the need for methodological restraint.