Generate accurate, unique meta descriptions

[LukasWallrich]

Why this needs addressing

The shared head template hard-cuts every description with printf "%.254s", which can split words. The audit also found generic fallback descriptions repeated across author and tag pages. The live author page /authors/kimberly-quinn/ currently uses the site-wide description and is marked noindex, follow.

Google does not require a fixed meta-description length; it recommends accurate, page-specific, useful descriptions and truncates snippets as needed. The issue is therefore duplication and poor generation, not the number 254 by itself.

Relevant implementation: layouts/partials/site_head.html.

Requested change

• Generate unique, human-readable descriptions for glossary terms, curated resources, author profiles, and tag archives from their actual content/metadata.
• When a fallback truncation is needed, truncate on a word boundary and treat the target as a presentation heuristic, not an SEO rule.
• Escape entities exactly once so literal “-style output is not emitted.
• Keep deliberate noindex behaviour for genuinely empty taxonomies/profiles, but do not use the generic site description as their page description.

Acceptance criteria

• Representative descriptions are page-specific, valid HTML attributes, and do not end mid-word.
• Duplicate-description counts for author/tag/resource samples are reduced or explicitly justified.
• Non-English descriptions render the intended characters once.
• Add a build-time check for missing or unintentionally duplicated descriptions on indexable pages.

Reference: https://developers.google.com/search/docs/appearance/snippet

[LukasWallrich]

Generation-pipeline clarification

For glossary descriptions, do not hand-edit generated term Markdown. Fix description generation in the shared Hugo head/template and, if additional generated metadata is needed, update _create_glossaries.py or its authoritative source data.