From Punch Cards to Prompt Engineering: The MHDBDB and the Future of Semantic Annotation with LLMs

This ACDH lecture by Katharina Zeppezauer-Wachauer and Julia Hintersteiner focuses on the technological redesign of the Middle High German Conceptual Database (MHDBDB), a long-running research resource for medieval German texts. Moving from relational databases and RDF-based graph architectures towards a TEI-first model, the lecture discusses the challenges of maintaining complex semantic infrastructures and the growing role of Large Language Models in annotation, search and interpretation. It demonstrates how TEI-XML can combine transparent and interoperable data structures with LLM-supported workflows while keeping scholarly expertise at the centre of the research process.

Learning Outcomes

After watching this video, learners will be able to:

• describe the technological evolution of the MHDBDB and the reasons for adopting a TEI-first approach
• explain how TEI-XML can support LLM-based and Expert-in-the-Loop research workflows
• discuss the opportunities and challenges of integrating Large Language Models into semantic annotation

ACDH Lecture