How the library is built

Methodology

The source documents are TEI P5 XML normalised to Unicode NFC. The publication pipeline preserves document divisions, textual blocks, editorial metadata and stable CTS identifiers.

Citation reconstruction

Corpus release 2.1.0 reconstructs only citation levels supported by the source structure and crosswalk. Passages that cannot be aligned safely remain explicitly local rather than receiving conjectural canonical references.

Presentation

The browser reader assigns layouts for prose, verse, drama, scripture, lexica, commentary, mathematical texts, fragments and papyri. Technical container labels are separated from authored headings and canonical references.

Reproducibility

Every generated work page links its machine-readable metadata, CTS URN, digital source when supplied, corpus release and revision date. The transformation does not create biographies, translations or summaries that are absent from the source.