ScriptNet - Text recognition of historical documents
Published May 19, 2025 | Version v1
Book Open

Praeteritum transcriptum. A Transkribus Tribute: Celebrating our First Five Years as a Cooperative (2019-2024)

Creators

  • 1. EDMO icon University of Twente
  • 2. READ-COOP SCE
  • 3. ROR icon University of Bern
  • 4. ROR icon University of Edinburgh

Contributors

  • 1. Universität Bern
  • 2. ROR icon University of Bern
  • 3. ROR icon Adam Mickiewicz University in Poznań
  • 4. British Library
  • 5. ROR icon Heidelberg University
  • 6. Katholieke Universiteit Leuven Faculteit Theologie en Religiewetenschappen
  • 7. ROR icon Royal Netherlands Academy of Arts and Sciences
  • 8. ROR icon National Archives of Sweden
  • 9. ROR icon University of Passau
  • 10. ROR icon University of Amsterdam
  • 11. ROR icon Noord-Hollands Archief
  • 12. Erasmus Universiteit Rotterdam Rotterdam School of Management
  • 13. ROR icon Instituut voor de Nederlandse Taal
  • 14. Dutch Language Institute
  • 15. ROR icon Bibliotheca Hertziana – Max Planck Institute for Art History
  • 16. ROR icon New York University Abu Dhabi

Description

Praeteritum transcriptum: A Transkribus Tribute

This publication commemorates READ-COOP SCE's first five years (2019-2024), examining how the cooperative structure behind Transkribus has transformed digital humanities and historical scholarship. Beyond celebrating technological achievements in automated text recognition (HTR/OCR), the work analyzes how a stakeholder-owned cooperative model offers a sustainable alternative to commercial platforms for cultural heritage preservation.

The tribute documents Transkribus' ecosystem as co-created by diverse stakeholders—software engineers, archivists, historians, and librarians—who serve as contributors, owners, and decision-makers rather than mere users. Through international case studies spanning various languages, scripts, and disciplines, it demonstrates the platform's adaptability to diverse scholarly requirements.

Central to the analysis is the productive symbiosis between tool developers and humanities scholars, where pooled user data has enabled the development of comprehensive historical text recognition models. The publication positions Transkribus as not merely a technological solution but as an ethical framework for AI development grounded in collective governance, transparency, and scholarly collaboration, ultimately redefining approaches to our shared written heritage.

Files

PraeteritumTranscriptum.pdf

Files (55.3 MB)

Name Size Download all
md5:6eb9e0b195600a6ca187240ba206c677
55.3 MB Preview Download

Additional details

Dates

Issued
2025-05-19