One of the most fundamental facets of preserving born-digital cultural materials is understanding the data formats in our collections. In contrast to the relatively constrained worlds of metadata and digitised materials, born-digital items can come in a huge range of formats, making format identification a crucial step towards understanding the informational and software dependencies we need to capture to make future access possible. How can today’s available file format identification tools and strategies help us understand the formats in our collections, and where are we still lacking ways to gather this critical information?

As part of the Registries of Good Practice project (a collaboration between the Digital Preservation Coalition and Yale University Library), we have been analysing and comparing data from a wide range of format identification tools and registries to better understand the current landscape. Combining this with methods drawn from ecological studies of species diversity, we have used the gaps between format registries to estimate the true scale of the format problem. Building on this evidence, we outline how a more distributed and responsive strategy would help us cope with the challenge of preserving our born-digital heritage.

In this paper I provide a practical rapid-assessment methodology for determining whether a preservation intervention is recommended based upon a prioritized, sequential list of positive and negative file format obsolescence risk factors.

Rather than a complex scoring system, the core output of this methodology is a simple set of conditions that aim to suit general use-cases for quickly determining whether a given format is at immediate risk of obsolescence and therefore whether an immediate preservation intervention is required.

Ensuring metadata compliance with the submission guidelines of preservation service providers is a labor-intensive process, requiring complex manual mapping and transformation. In this paper, we present a work-in-progress prototype for the automated preservation of digital objects from cultural heritage collections. To achieve this, we developed the Aggregator - a tool that manages preservation key functions such as SIP compilation, object upload, data integration, package handling, access control and client management. Designed for multi-client use, the system allows the preservation process to be entirely controlled by the client. As a result, no additional personnel resources are required from the digital preservation service provider for ingest, access, or package handling.

Tools and services that are both reliable and approachable to the end user benefit everyone in digital preservation. From validating digital objects to creating a submission information package, the goal is to enable a dependable and streamlined process. We at the National Digital Preservation Services in Finland have developed tools to help with this, two of which will be covered in this paper. The File Validation Service is a stand-alone online file validator that helps in catching errors in files as early as possible in the life cycle of digital content. The Pre-Ingest Library is a Python library that automates the creation of complex metadata documents needed for ingesting content into the services.

This paper examines the challenges and methodologies associated with the long-term preservation of Prompt resources within the domain of artificial intelligence, and proposes a storage framework. The article delves into the dynamic dependencies, as well as the multimodal characteristics encompassing text, images, and videos associated with Prompt. It develops a five-layer storage model comprising: an infrastructure layer (utilizing Docker and hybrid cloud solutions), a data storage layer (incorporating relational and vector databases), a embedding layer (focusing on multimodal embedding), and an application layer (encompassing OAIS processes and API services). The study identifies several technical challenges, including the optimization of storage costs, the safeguarding of privacy, and ensuring version compatibility. Furthermore, it outlines prospective applications in the realm of large medical models, with the objective of providing a systematic solution that ensures the integrity, accessibility, and compliance of high-value Prompt resources.