Method & SourcesMétodo y fuentes

This edition organizes United States newspaper coverage of the Dominican Republic from the Library of Congress Chronicling America archive. The curated timeline holds 1418 records from 352 newspaper titles, 1820 to 1963, in chronological order.

Esta edición organiza la cobertura de la República Dominicana en periódicos estadounidenses procedente del archivo Chronicling America de la Biblioteca del Congreso. La cronología curada contiene 1418 registros de 352 periódicos, de 1820 a 1963, en orden cronológico.

Article texts. Full texts are vision transcriptions of the Library of Congress page images: the article only, no neighbouring items. The process has known limits, measured in this edition's adversarial review (2026-11-01): a printed headline repeated at the start of a text is stripped at build; transcriptions that fail an automated containment check against their page are flagged or held on their pages; passages the model could not read are supposed to be marked [illegible], and that discipline is spot-audited because early output under-used the marker. 106 of 1418 records carry full text in this pilot. Model self-reported confidence is not published; what is published is the QA state.

Textos de los artículos. Los textos completos son transcripciones visuales de las imágenes de página de la Biblioteca del Congreso: solo el artículo, sin textos vecinos. El proceso tiene límites conocidos, medidos en la revisión adversarial de esta edición (2026-11-01): un titular impreso repetido al inicio se elimina al construir; las transcripciones que no superan una verificación automática contra su página se marcan o se retienen en sus páginas. 106 de 1418 registros llevan texto completo en este piloto.

Passage summaries. Records not yet extracted carry a summary generated from the located OCR passage - not from the full article. Passage windows can cross article boundaries; summaries are labeled "Passage summary" wherever they appear, and any summary that fails a fidelity check is withheld until the article itself is extracted. The located passage is displayed in full on every article page, and the excerpt digest printed in Provenance is computed over exactly that displayed text, so any reader can recompute it.

Resúmenes del pasaje. Los registros aún no extraídos llevan un resumen generado desde el pasaje OCR localizado, no desde el artículo completo. Los resúmenes se etiquetan «Resumen del pasaje» dondequiera que aparezcan, y todo resumen que no supere una verificación de fidelidad se retiene hasta extraer el artículo. El pasaje localizado se muestra completo en cada página, y el resumen SHA-256 de Procedencia se calcula sobre exactamente ese texto mostrado.

Inclusion rule. A record belongs to this timeline when the located item substantially concerns the Dominican Republic (including the island's earlier states and the name Santo Domingo), its government, people, territory, or foreign relations. Items primarily about Haiti are included when the item itself substantially engages Santo Domingo or Hispaniola-wide affairs; a passing mention is not enough. In nineteenth-century usage "Dominica" frequently names the Dominican Republic itself and is judged in context; the island of Dominica is a different place and is excluded. The collection was built by a keyword screen over the Library's OCR followed by a scored curation cut; the screen's measured precision is published in the review, the grey pool is quarantined rather than imported, and borderline records carry review flags rather than silent verdicts.

Regla de inclusión. Un registro pertenece a esta cronología cuando el texto localizado trata sustancialmente de la República Dominicana (incluidos sus estados anteriores y el nombre Santo Domingo), su gobierno, su gente, su territorio o sus relaciones exteriores. Los textos principalmente sobre Haití se incluyen cuando el propio texto aborda sustancialmente Santo Domingo o los asuntos de toda la isla; una mención pasajera no basta. En el uso del siglo XIX, «Dominica» nombra con frecuencia a la propia República Dominicana y se juzga en contexto; la isla de Dominica es otro lugar y se excluye.

Evidence labels. Claims on this site are labeled by basis: Documented for material substantiated by a cited primary source, and Oral tradition / loosely verified for family or oral accounts that may or may not be accurate, spelled out for the high-scrutiny reader.

Etiquetas de evidencia. Las afirmaciones de este sitio se etiquetan según su base: documentado, para material sustentado por una fuente primaria citada, y tradición oral / verificación parcial, para relatos familiares u orales que pueden o no ser exactos.

Provenance. Every article page carries its record id and SHA-256 digests, and links to the Library of Congress page. Source URLs in this collection were derived from the Library's addressing scheme and are being verified by a standing link sweep; a record whose URL fails is flagged on its page. In the raw archive, file names are the SHA-256 of the source URL (an addressing convention, not a content hash); the content digest is the source-text SHA-256 printed on each page. This edition, its data files, this review's evidence, and the raw OCR archive form the preservation package, snapshotted with a manifest outside the code repository.