遇见数据集

Registrum B of Hernando Colón: AI-assisted reading of the 1905 facsimile (working data v3.7, unverified)

收藏
Zenodo2026-10-29 更新2026-11-05 收录
官方服务:

资源简介:

AI-assisted reading of the entries of Registrum B, the catalogue of the library of Hernando Colón (1488-1539), taken from the 1905 facsimile (A. M. Huntington, Catalogue of the Library of Ferdinand Columbus, New York, 1905; digitised by Google via Internet Archive; public domain (US)). Version v3.7 replaces v3.5 (10.5281/zenodo.23736838), which remains available. The table has one row per box entry and gives the box number, diplomatic text, author, title, format, place, year and other numbers, with flags for doubtful numbers. The same data are also given as TEI, SQLite and an offline search page, together with a table of the 726 rows that differ from v3.5 and a list of unresolved rows. STATUS: working data, not a critical edition. The pages were read by language models, and human verification has not been performed. The readings are unverified and must not be quoted as the text of the manuscript. The accuracy of author, title, format, place and year has not been measured. What changed from v3.5: a whole-column shift of box numbers was found on 61 page-columns (each number sat one entry too early) and corrected from the scan readings. 506 rows were renumbered, 61 continuation rows merged, 52 rows added from the scan with only their first words, and 2 numbers inferred from sequence. 4,205 of the 4,231 distinct box numbers are now present (99.4%); 26 are still missing (4052 and 3683-3707). 68 rows contain only the first words of the entry and have no author or title yet. Coverage means only that a number was found. It says nothing about whether the reading of an entry is correct. Limits: 49 of the 61 corrected columns were read by one model pass only; flagged and unresolved rows are marked in the table; the TEI file is well-formed but was not validated against the official TEI schema. The README and the changelog list the limits in full. Human checking is still to come, and its results will be published in a later version, whatever they show. The deposit contains text only. No page images are included. No data from the Concordado, from the website of the Biblioteca Colombina, or from CERL were used. This is an independent project, not affiliated with any institutional research project on these manuscripts. It makes no claim to be a critical edition and no claim about the identification of the printed editions or surviving copies recorded in the catalogue.

提供机构:
Zenodo
创建时间:
2026-10-29
二维码
社区交流群
二维码
科研交流群
商业服务