Blog content collection design

Decision

• Manage blog posts as Markdown/MDX files under /blog.
• Use the following frontmatter schema:
  • title (required, string)
  • description (optional, string)
  • publishedAt (required, date)
  • updatedAt (optional, date)
  • draft (optional, boolean, default false)
  • tags (optional, string[])
  • image (optional, string)
  • noindex (optional, boolean, default false)
  • lang (optional, string, default ja)
  • slug (optional, string, derived from filename when omitted)
  • prev (optional, true | false | string, explicitly control previous link)
  • next (optional, true | false | string, explicitly control next link)
• Follow the i18n URL strategy by deriving URLs from the lang + slug combination. Use the same slug across translations.
• Sort lists by publishedAt and use updatedAt for “last updated” display.
• Treat draft as the flag for excluding posts from listings/feeds/build targets.
• Treat noindex as metadata indicating the post should be excluded from search engine indexing.

Context

A consistent schema is required to reliably generate listings, detail pages, and SEO metadata for long-term blog operations. The design must align with the i18n approach (Japanese as default, additional languages under subdirectories).

Options

• Markdown/MDX with Content Collection
• External headless CMS
• Manual JSON/TS management

Rationale

• Markdown/MDX keeps authoring and review lightweight.
• Content Collection schemas provide build-time validation.
• Explicit lang and slug fields maintain consistent URLs aligned with the i18n strategy.

Consequences

• All blog posts must follow the schema above.
• Operational rules for draft and noindex need to be shared as publication criteria.
• Adding languages requires creating translations with matching slug values.