What is "Preservation Metadata"?

Question

I often see 'preservation metadata' used in sentences which suggest it is distinct from other types of metadata, e.g. distinct from 'descriptive', 'structural', 'rights' or 'technical' metadata. However, it would seem that these "other" forms of metadata also have a part to play in preserving digital material, for example technical metadata about an image's dimensions could be used within Quality Assurance processes.

So, is 'preservation metadata' a distinct class/set of metadata? And if so, what does that set constitute?

Or should 'preservation metadata' be used to cover all metadata useful for preserving material, irrespective of whether it is 'descriptive', 'techincal', etc.?

Answer 1

I would point your attention to a wonderful report published at DPC last year on the topic of preservation metadata. It can be found here:

http://www.dpconline.org/component/docman/doc_download/894-dpctw13-03

Comments

As a follow up, the report states on page 2:

"Preservation metadata is metadata that supports the distinct requirements of digital preservation: maintaining the availability, identity, persistence, renderability, understandability and authenticity of digital objects over long periods of time."

Though not meant to be definitive--what can be definitive without an official standard?--I think this understanding is an adequate start.

Answer 2

I do not think the term is used consistently. I think some people mean PREMIS, some METS, some PREMIS-in-METS, while others mean "metadata that has no parallel for non-digital material and only turns up for digital items". I'd imagine the only common thread would be fixity values - we'd all agree that hashes/checksums/signatures are required to preserve digital resources.

I'd be interested to know if non-digital materials have something they call 'preservation metadata' and whether it's more tightly defined.

Answer 3

No "type" of metadata is a distinct subset of metadata. Take these three scenarios.

• Copyright metadata restricting the viewing of a video.
• Location metadata recording the location of an image referenced in an HTML file.
• Format metadata describing the version of software used to create a CAD file.

Each of these pieces of metadata have been categorized both as preservation metadata and another type of metadata (rights management, structural, and technical) by the community at different points. That's not to say that preservation metadata encompasses all metadata. I'm sure there are valid arguments that it might be. Similarly, there are valid arguments that descriptive metadata, technical metadata about digitization, or other metadata might or might not be required for preservation.

Each "type" of metadata is most useful as an internal shorthand what that type of metadata has been defined to be. Outside of a specific organization or context, don't trust that you are speaking about the same metadata elements just because you're discussing the same "type."

If you are thinking about what defines preservation metadata for your context, I like the NDSA's definition. "The contextual information necessary to carry out, document, and evaluate the processes that support the long-term retention and accessibility of digital content."