Jump to content

Applications and code standards

From Metopedia


This page sets standards for applications, scripts, code, datasets, and computational resources published on Metopedia.

Metopedia may publish code and applications as research materials. Code must be treated as part of the evidence and method record, not as an unquestionable result.

Core rule

Code published on Metopedia should be readable, documented where necessary, licensed where possible, reproducible, and safe for research use. Outputs must be interpreted through method and evidence.

Required information

A code or application page should include:

  • title;
  • purpose;
  • author or maintainer;
  • version;
  • date;
  • license;
  • dependencies;
  • supported environment;
  • inputs;
  • outputs;
  • example use;
  • test data;
  • known limits;
  • known failure cases;
  • security cautions;
  • related article or dataset;
  • repository or archive link where available.

Code integrity

Where practical, code releases should provide:

  • version number;
  • checksum or hash;
  • change notes;
  • dependency versions;
  • reproducible example output;
  • source archive;
  • known issues.

Data integrity

Datasets should identify:

  • source;
  • collection method;
  • date range;
  • transformations;
  • omissions;
  • filtering rules;
  • deduplication rules;
  • privacy handling;
  • license or access limits;
  • known bias or incompleteness.

Security

Metopedia does not permit malicious code, credential theft, spyware, destructive scripts, unauthorized access tools, malware, or instructions whose primary purpose is abuse. Code that touches files, networks, credentials, APIs, user accounts, or external services must clearly describe risks and expected behavior.

Analytical claims based on code

When an article relies on code output, the article should state:

  • what code was used;
  • what version was used;
  • what data was used;
  • what preprocessing was performed;
  • what assumptions were made;
  • how the output was checked;
  • what the output supports;
  • what the output does not prove.

See also