Risk Assessment

Generated: 2026-07-22T15:10:59+00:00

Rejected Alternatives

• Discipline-first folders as the primary architecture
• Project-only organization without shared concept ownership
• Docs-only static site generators as the full platform model
• Manual curation of registries and backlinks

Architectural Tradeoffs

• A hybrid concept-plus-project model is more complex than a plain docs tree, but it preserves cross-project reasoning.
• Hybrid retrieval requires more infrastructure than local full-text search, but single-mode search will fail at scale.
• Typed metadata adds authoring overhead, but the repository is already large enough that untyped growth would create long-term entropy.

Scalability Risks

• Flat intake directories become unmanageable well before 10,000 documents.
• Semantic indexing costs increase sharply if chunking and canonical summaries are not normalized.
• Graph density can become noisy without controlled relationship vocabularies.

Maintenance Costs

• Metadata linting and schema evolution must be treated as first-class engineering work.
• Canonical concept ownership requires editorial governance.
• Search quality tuning needs periodic evaluation against real queries.

Technical Debt To Avoid

• Hard-coding navigation in website templates
• Allowing duplicate status vocabularies
• Embedding relationships only in prose
• Using filenames as the only durable identifiers

Future Migration Concerns

• If Astro no longer fits, keep generated data products framework-agnostic so the site can be replatformed without reclassifying the corpus.
• Preserve immutable manifests and graph exports so future researchers can reconstruct repository state independent of the website stack.