From source to screen.

Finding and identifying works

The wider project uses authority records, library catalogues, historical newspapers, inspected collections and established bibliographies. Titles and versions are reconciled before a work is treated as one canonical record. Availability alone is never evidence of accuracy or permission.

Choosing and transcribing a text

A named copy-text is selected from inspected evidence. The transcription is diplomatic: lineation, stanza divisions, spelling and punctuation are preserved. Technical normalisation may support search or display, but it does not rewrite the canonical poem.

Rights and publication

Publication is deny-by-default. Each emitted text needs an affirmative Australian work-level rights determination and the required editorial, textual and cultural clearances. Unresolved, held and protected records remain outside the public collection.

Context and uncertainty

Sourced fact is separated from interpretation. Missing evidence is labelled or omitted rather than guessed, and historical or cultural sensitivity can trigger specialist review even where copyright has expired.

Use of AI

AI may assist controlled research, comparison, drafting and technical work. It is not treated as a source, and it does not replace inspection of evidence, editorial judgment or recorded review.