WILL AM-FM-TV/University of Illinois at Urbana-Champaign
& Aggregating Metadata without Massive Transformations

Context
PBCore Integration
Contacts

Other Case Examples

Recently, the University of Illinois at Urbana-Champaign conducted a campus-wide Metadata Ecology Project. Anyone producing content, including video/tv and audio/radio media, were included in the interview process. The project's mission was to discover what kind of information was being generated, how it was used and saved, and how data was shared. The conclusion was that everyone "did it differently." Few did it adequately. However everyone involved thought information management and commonly available data management tools would be tremendously helpful and would not be considered a totally abstract concept.

It was determined that many different groups on the University of Illinois campus were extremely interested in different knowledge domains, i.e., scientific, historical, public media, news stories, etc. The campus had already embraced the Open Archives Initiative (OAI) for data sharing and harvesting, thus allowing one group to extract content from other knowledge domains.

Within WILL, the Metadata Ecology Project catalyzed the quest to determine how best to aggregate content and to integrate the generation and flow of data across the many different systems within their operations (including production, web publishing, automation and playout, as well as traffic and scheduling). A System Integration position has since been created to pursue the integration goals. However, for the present, rather than tackle all of these integration tasks at once, WILL decided to start with a proof-of-concept in which key elements of data exchange could be developed, tested, and demonstrated.

The target for information re-development was the WILL website, http://www.will.uiuc.edu/. The method was to implement Content Management. Instead of focusing on actual tools for cataloging media descriptions, WILL decided to use website content managment as the cataloging tool. Their interest in metadata was to recirculate vetted metadata descriptions out to the web in a structured way, avoiding the construction of numerous data transformation bridges. Consequently, WILL developed a data model around their Prairie Fire series that for over 15 years had told stories about the history, culture, and events within central Illinois.

The task was to account for all metdata elements that would need to be expressed on the Prairie Fire web site (contributors, image, location for media file, bit rate, all technical and format data). It was determined that what didn't need to be published directly on web site for visitor consumption could be expressed in an XML record. As part of the internal/intranet functions of the content management system, producers were asked to submit program and segment metadata via a web-based form. As a sidebar consideration, WILL has found that the use of folksonomies in tagging or describing assets adds value to the formal metadata dictionaries and schemas that are employed; there is enough evidence from WILL's perspective to indicate that folksonomy descriptors are useful, worthwhile, and ought to be facilitated and actively harvested, thus becoming part of a metadata record that evolves over time.

Another conclusion WILL has drawn is that an organization should try not to "overthink" the problem or the solution with regard to data collection, aggregation and exchange. Actual cataloging tools are almost irrelevant when the true focus is placed on the metadata dictionaries and schemas in use. As long as metadata can be reshaped or exported into compliant metadata documents using XML Schema Definitions (XSD), data sharing can be facilitated with the use of many tools from Excel spreadsheets, FileMaker databases, or content management systems.

As WILL embarked on its data aggregation and interoperation discovery process, they began thinking about metadata for public media, then became aware of PBCore as a metadata dictionary and XSD for public broadcasting. PBCore offered ready-made authority references, picklists of acceptable vocabulary terms, and rules for structuring the data for individual metadata descriptions, in addition to "style sheets" for data exchange. WILL saw PBCore as the public broadcasting version of the Open Archive Initiative (OAI) for data sharing and harvesting.

In its proof-of-concept work using a content management system for their Prairie Fire website, metadata of interest to the general viewing public was pushed from the central database (populated by producers using web-based forms on the WILL intranet) and published to the site's web pages. More detailed metadata was made available as XML documents, generated from the PBCore XML Schema Definition.

Each media item presented on the Prairie Fire website contains a link (lower right) to a PBCore XML metadata record that presents itself as an XML document within the web browser. A sample is presented below.

WILL admits that this prototype XML record is likely of diminished interest to the general public. But as a proof-of-concept for the ability to generate XML metadata records compliant with a metadata schema according to its XML Schema Definition (XSD), the implementation has been very informative and enlightening.

CONTACTS:

Other Case Examples