Key points are not available for this paper at this time.
Emulation and migration are still our main tools for digital curation and preservation practice. Both strategies have been discussed extensively and have been demonstrated to be effective and applicable in various scenarios. Discussions have primarily centered on technical feasibility, workflow integration, and usability. However, there remains one important aspect when discussing these two techniques: managing and preserving operational knowledge. Both approaches require specialized knowledge but especially emulation requires future users to also have a great variety of knowledge about past software and computer systems for successful operation. We investigate how this knowledge can be stored and utilized, and to what extent it can be rendered machine-actionable, using modern large language models. We demonstrate a proof-of-concept implementation that operates an emulated software environment through natural language.
Rechert et al. (Mon,) studied this question.