About This Collection — How the Data Came to Be

A Plain Word Up Front

This collection is an act of curation, not scholarship. It was assembled by gathering, organizing, and cross-checking information that is already publicly available — from library catalogs, reference works, and digitized texts — with the help of software tools and a large language model. It does not come from personal expertise. I’m not a credentialed historian, theologian, or bibliographer, and this project didn’t begin with academic credentials — it began with curiosity and a commitment to getting it right.

What you’re seeing is collective knowledge — the accumulated record of many libraries, editors, and scholars — gathered into one place and made easy to explore. That carries real value, and it carries real limits. Both are described honestly below.

This is not yet an academic-grade resource. It’s a careful, good-faith starting point that we are steadily working to improve. Where we’re confident, we say so; where authorship or dates are uncertain, we try to flag it rather than hide it.

Where the Information Comes From

Names, dates, identities, and works are drawn from established public reference sources and verified against more than one wherever possible.

Identity & authority records — who a person was, and a stable way to identify them across catalogs:

Texts, editions, and bibliography — what was written, and where to read it:

Reader interest — used only to decide what to work on first, not who is “important”: aggregate Wikipedia pageview data.

How the Biographies, Descriptions, and Summaries Are Written

The written prose on this site — author biographies, work descriptions, and short summaries — is drafted with the assistance of an AI language model, specifically Claude, made by Anthropic. The model is asked to write from an author’s or work’s known facts (dates, tradition, titles), and a person then reviews the draft before it’s saved.

We want to be completely clear about what this does and doesn’t mean:

  • The descriptions are AI-assisted drafts reviewed by a human, not original research and not peer-reviewed scholarship.
  • AI models can sound confident and still be wrong — they can misattribute, misdate, or invent plausible-sounding detail. Review reduces this; it does not eliminate it.
  • Where a description makes a claim that matters, we try to ground it in the reference sources above rather than the model’s unaided memory.

If a biography or summary reads smoothly, that’s the writing tool doing its job — it is not a signal that the underlying facts have been academically verified.

How the Collection Is Curated and Checked

The data lives in a custom catalog with editing tools built for this project. In practice, curating a record means:

  • Establishing identity — matching a person to authority records (VIAF / Library of Congress) so we’re confident who we mean.
  • Confirming the body of work — listing what an author actually wrote, and marking how securely each work is attributed to them. Authorship is graded: accepted (confirmed by a reviewer), traditional (the common, long-standing attribution), disputed, or spurious (historically attributed but rejected by scholarship). Doubtful and spurious works are kept for reference but set apart.
  • Linking to where you can read it — free public-domain texts where they exist, and editions to purchase where they don’t.
  • Cross-checking — comparing the above against more than one source before a record is treated as settled.

Software (including AI) assists every step — suggesting works, proposing identifiers, drafting prose — but suggestions are treated as leads to verify, not facts to publish.

What This Means — and Its Limits

In the spirit of being useful and honest:

  • Collective knowledge can be wrong. Catalogs disagree; older sources repeat old errors; popular accounts simplify. Aggregating many sources helps, but it inherits their mistakes.
  • Ancient and medieval authorship is genuinely uncertain. Many early works are “attributed to” a figure rather than provably theirs. We flag this rather than pretend to certainty.
  • AI assistance introduces its own risk of fluent-but-wrong text, which is why prose is reviewed and important claims are tied back to sources.
  • Coverage is uneven. Some authors are richly documented; others are thin stubs awaiting work.
  • This is a living project. Records are added and corrected continually, so what you see today may be refined tomorrow.

Treat this site as a well-organized doorway into the Christian written tradition — a place to discover, orient, and find the actual texts — and verify anything you intend to rely on against the primary sources linked throughout.

Where We’re Headed

The goal is to keep raising the bar: deeper sourcing, clearer citations, more records confirmed against authoritative scholarship, and steadily fewer that rest on “traditional” attribution alone. We’re not an academic resource yet — but that’s the direction we’re working in, deliberately and in the open.

Found Something Wrong?

Corrections are genuinely welcome and make the collection better for everyone. If you spot an error — a wrong date, a misattributed work, a questionable claim — please tell us at [CORRECTIONS EMAIL coming soon]. Pointers to a better source are especially appreciated.