Understanding metadata
The third article in a series of insights about metadata.
We’ve talked a lot about language, but language is just one, albeit a big one, of the metadata elements we use to manage information objects. In fact there is a whole range of terms and Identifiers which make up what we refer to as metadata.
Many different types of metadata may be developed for any given piece of information. Metadata can be used for administrative control, security, recognition of personal information, content rating, rights management and preservation. Broadly speaking:
Descriptive metadata is data added to content to make it easier to find;
Administrative Metadata – makes content easier to manage; and
Structural metadata makes content easier to use as it provides information about how content is to be read, manipulated and displayed.
The usability and effectiveness of any information system is clearly affected by the quality of its metadata.
If you have built a digital music library by uploading hundreds of CDs you will know what I mean. Not only will you load up the tunes (content) but also the metadata describing each tune. Your new metadata library provides evidence of what happens in the absence of standards for describing albums, genres, composers etc.
You will find duplications, overlaps, different syntax, misspellings or no metadata at all. If the metadata for specifying the play order is missing, you may find yourself listening to the last movement of Beethoven’s 9th first.
Of course the record producers of earlier generations weren’t able to anticipate that decades later we would be browsing for our music using metadata. The I-gadget provides the perfect example of how our need for metadata has escalated with digitalisation.
Metadata standards are requirements that are intended to establish a common understanding of the meaning or semantics of the data, to ensure correct and proper use and interpretation of the data by its owners and users.
There are literally thousands of metadata standards covering recordkeeping, web resources, geographic data, the arts, scientific data etc. To get an idea of the scope and variety check out this Visualization of the Metadata Universe by Indiana University Libraries.
So many standards to choose from, so it’s essential to understand the types of metadata standards and their purpose:
Type | Example |
Data structure standards (metadata element sets, schemas).These are “categories” or “containers” of data that make up a record or other information object. | Dublin Core Metadata Element Set (DCMES) The set of MARC (Machine-Readable Cataloging format) fields Encoded Archival Description (EAD) |
Data value standards (controlled vocabularies, thesauri, controlled lists). These are the terms, names, and other values that are used to populate data structure standards or metadata element sets | Library of Congress Subject Headings (LCSH), Art & Architecture Thesaurus (AAT), Getty Thesaurus of Geographic Names (TGN), Keyword AAA Thesaurus |
Data format/technical interchange standards (metadata standards expressed in machine-readable form). This type of standard is often a manifestation of a particular data structure standard (type 1 above), encoded or marked up for machine processing. | MARC21, EAD XML DTD, Simple Dublin Core XML schema |
Data content standards (cataloging rules and codes).These are guidelines for the format and syntax of the data values that are used to populate metadata elements.. | Anglo-American Cataloguing Rules (AACR), Resource Description and Access (RDA), International Standard Bibliographic Description (ISBD, Describing Archives: A Content Standard (DACS) |
A Typology of Data Standards
Source: Introduction to Metadata 2ndEd., Getty Publications
Conformance with standards confers authority and builds the rigour into enterprise metadata. In the absence of external standards/specifications, organisations should look to establishing their own enterprise standards. Furthermore in any organisation, there will be a need to conform with multiple standards.
To quote from ISO 23081 Metadata for Records:
Different perspectives on metadata are possible and can coexist.
It pays to bear that in mind when designing and developing information systems.