Thursday, March 7, 2013

Class No. 7: A Beginner's Guide to Metadata

I recently took my seventh DAS course: A Beginner's Guide to Metadata, taught by Greg Colati and Jessica Branco Colati. This was another webinar course, recorded live back in 2008. I was a little concerned that the information would be too basic and, after five years, slightly out of date, but overall I felt that it was a well-organized, informative presentation. The goals were to provide a basic overview of metadata - what it is, where it comes from, what its components are, and how to choose the "right" schema(s) for one's own organization.

Metadata has of course existed for thousands of years in the form of analog inventories and cataloging systems, but the word "metadata" comes to us from the IT world (incidentally, the capitalized term "Metadata" is actually a registered trademark of the Metadata Company - who knew?). Metadata was originally considered "cataloging for geeks," "cataloging for guys," or just "cataloging for non-librarians," but over time it has become firmly entrenched in the lexicon of archivists and librarians. Simply defined, it is a tool for the identification, management,and use of information resources. It must be structured according to a schema, it must describe an information resource, and it should be useable by both machines and humans. Once created, it is never perfect or truly complete; it must be continually improved upon and eternally maintained. Put another way, metadata is language, made up of syntax, structure, and semantics. As archivists we should aim to be "multilingual metadata speakers."

Having defined metadata, the instructors described the various "typologies" of metadata, such as primary, secondary, or tertiary; descriptive, administrative, technical, or structural; global, community, or local; and embedded or associated. They then explored the idea of metadata as language by picking apart a few lines of an EAD finding aid. The syntax for this particular piece of code was XML, which governed the fact that some information was in brackets while other information was not. The structure was EAD (Encoded Archival Description), which governed the specific elements that could be used, and the hierarchical order in which they appeared. The semantics were governed by DACS (Describing Archives: A Content Standard), which informed the content of the finding aid - the way the name of the creator was formed, for example, or the format of the collection title. The instructors noted that the syntax and structure of our metadata will likely change over time, but hopefully the semantics will remain relatively stable. I thought this was a helpful exercise, in the same way that the basic rules and components of grammar must be understood more explicitly when we attempt to learn a foreign language.

The last part of the course covered general things to consider when choosing a metadata schema (or schemas) for your own institution. This decision requires four steps:
  1. Identify your needs. What kinds of objects do you have, and what kind of information do you need to collect about them? What do you want to be able to do with your metadata? How does your audience expect to be able to find and interact with your holdings?
  2. Identify your resources. Creating quality metadata is costly, but keep in mind that tasks left for sometime in the distant future will likely never get done. Do you want to spend your resources providing a high level of access to a few things, or a low level of access to a lot of things? If your institution follows MPLP, can you justify item level metadata for digital objects? And how do you define an "item"? Does user-created metadata have a place in our catalogs? It may be free to obtain, but there are costs involved in monitoring it for quality control and accuracy.
  3. Test your vernacular. Is the schema applicable? Is it useful? Does it meet the needs of your primary audience? Can you successfully communicate with it?
  4. Optimize your efforts. Look at the interoperability, shareability, reusability, and "archivability" of the schema. Remember that when moving from one metadata language to another some pieces of information will translate directly, some will be aggregated with others into a broader term, and some will be lost altogether.
For archivists with a good understanding of traditional archival practice but with limited experience creating or interacting with encoded metadata, or for those who might like to revisit the basics, I think this webinar is a great starting point. As always, please feel free to comment if you have questions or feedback, and thank you for reading!