# About this project and how it is checked

This site surveys the ways of finding out whether software works — one page per
method, from types and static analysis through tests and fuzzing to monitoring
and mathematical proof.
[Methods](https://quality.stereobooster.com/methods.md) is the catalog, [the axes](https://quality.stereobooster.com/axes.md) the
classification behind it.

## How claims are checked

- **Citations are checked against a primary source.** A work is read, or its
  identity confirmed against a bibliographic index, before it is cited. Where the
  primary text could not be reached — a paywalled standard, a claim confirmed only
  from a secondary source — the entry records that instead of dropping the caveat.
- **Peer-reviewed work is the default source.** Journal and conference literature
  carries most of the evidence here. A claim resting on an industrial case
  study, a preprint, or a tool's own documentation is labeled as such rather
  than presented as a finding.
- **Evidence is labeled by strength.** A measured result is reported as a
  finding, thin support is marked as a plain note, and folklore is named as
  folklore — see
  [Testing folklore](https://quality.stereobooster.com/testing-folklore.md) and
  [Design folklore](https://quality.stereobooster.com/design-folklore.md).
- **Adoption is not evidence.** Popularity, longevity, and authority don't stand
  in for a measured result. Where a practice has no empirical support, the page
  says so rather than arguing from how many teams follow it.
- **Every method is described in the same frame.** The same axes and the same
  questions on every page, so two methods can be compared instead of each being
  described on its own terms.

## What this does not guarantee

**Working copy**

This is a working copy (preprint) — possibly containing errors. Pages vary
in maturity, and nothing here has been through review by a subject expert.

The author makes no claim to expertise in every area covered here. The aim is
narrower: to organize what's known and hold each claim to a stated check.

## Author

**stereobooster** has worked as a developer for 20 years, in fintech, travel,
marketing technology, and learning technology. Code is at
[github.com/stereobooster](https://github.com/stereobooster) and writing at
[stereobooster.com](https://stereobooster.com).

## Citing this work

Nothing here carries a version number, so a citation records the date the page
was read.

```bibtex
@misc{stereobooster2026,
  author = {stereobooster},
  title = {{Software Quality: a field guide to knowing whether your software works}},
  year = {2026},
  url = {https://quality.stereobooster.com/},
  note = {Working copy. Accessed <date>.}
}
```

A single page takes its own title and URL in place of the site's —
`title = {{Mutation testing}}`,
`url = {https://quality.stereobooster.com/mutation-testing/}`.

The content is CC BY-SA 4.0: attribution, a link to the license, and the same
license on anything derived from it.

## Corrections

A correction, a study that changes a conclusion, or a method missing from the
catalog belongs in the
[issue tracker](https://github.com/stereobooster/software-quality-skills/issues).
That repository holds the Claude Code skills distilled from this knowledge base;
its issues are the channel for the content here as well.

## Support this work

This knowledge base is unfunded and free to read. Support goes through
[thanks.dev](https://thanks.dev/u/gh/stereobooster).

## Referenced by

- [Software Quality](https://quality.stereobooster.com/index.md) · Overview