Direct answer
The site improves search coverage by mapping each important concept to one canonical page and a small set of supporting pages. It does not create thin doorway pages for every query variant. The goal is to make the best page easier to find, not to multiply weak pages.
Core topic clusters
| Entity | Canonical route | Supporting examples |
|---|---|---|
| Model breeding | /foundations/model-breeding | Recombination, merging, lineage, adapter registry, evaluation gates. |
| Teleodynamic AI | /foundations/teleodynamics | Viability mathematics, resource closure, no-op, metastability. |
| Digital Four Fs | /foundations/digital-four-fs-cycle | Four Fs, evolutionary Four Fs, code vs model breeding. |
| Small-model ecologies | /foundations/small-model-ecologies | Router, coalitions, browser runtime, local sovereignty. |
| Model merging | /benefits/model-merging-upside | Safe merge lab, adapter registry, recombination operators. |
| Plain PHP deployment | /operations/plain-php-package-contract | Root hosting, deployment contract, package checklist. |
Internal link rule
Every high-value page should link upward to its section index, sideways to related concepts, downward to implementation detail, and backward to source reports. The shared article template already adds breadcrumbs, section navigation, related cards, and source panels; authors should add contextual links inside the body when a concept is introduced.
Quality rule
A page is publishable when it answers a concrete reader question, teaches a useful system design decision, states limits, and links to evidence. A page is not publishable merely because it targets a keyword phrase.
Source reports used for this guide
These reports are preserved verbatim in the site archive. The guide above is an editorial synthesis and may narrow, qualify, or reorganize claims from the source material.