Attributes and Values Describing Entities Metadata At the most basic level, metadata is just another term for description, or information about an entity. In creating a descriptive schema, you are creating a template for assigning metadata to a particular group of items in a structured way. Attributes and values Attributes—also known as characteristics, properties, or elements—are the categories we use to describe a specific kind of entity more precisely. Values are a way to describe the possible contents of an attribute. Slide 3 Examples of attribute/value pairs These metadata elements for automobiles would seem familiar to many of us: Attribute Make Model Trim line Year Value The name of the company that produced the car (Toyota) The product line or brand (Camry) Model variation (SE) The year the car was produced, in YYYY format (2007) Slide 4 Examples of attribute/value pairs These metadata elements for wine would seem familiar to many of us: Attribute Value Country of origin The name of country where the wine was produced (France) Region of origin An area within the country of production (Bordeaux) Color The overall tint (red) Vintage The year wine was made, in YYYY format (2007) Slide 5 More examples of attribute/value pairs Attribute/value pairs that might enable more exploitative control for cars could be something like: Attribute Value Envy quotient How much will others be jealous of my sweet ride, on a scale of 1-10? Carelessness negation To what degree can one blow off regular maintenance without having a breakdown, on a scale of 1-10? Commuter fatigue How will you feel about the car after driving it 25 miles each day, in traffic (love car, hate car, indifferent to car, don’t know just need a drink) Slide 6 More examples of attribute/value pairs Attribute/value pairs that might enable more exploitative control for wine could be something like: Attribute Cuisines Friendliness Character Value With what styles of food will this wine go best? What percentage of guests at my family barbecue will like this wine? Which term best describes this wine (silky, bold, lush, succulent, opulent, zesty)? Slide 7 Schemas, or attribute sets A schema is a set of attributes and associated value parameters designed to describe a particular type of entity. Schemas may be encoded in a particular syntax for manipulation by people or computers. Schemas may also be associated with rules for creating records (that is, assigning attributes and values to specific resources). Slide 8 Dublin Core: a schema The Dublin Core is a metadata schema for describing (primarily) information resources. It includes a set of elements (attributes) and associated value parameters. A goal of Dublin Core is to provide a simple set of standard attributes that apply to most documents. By making it easy to comply with Dublin Core standards, interoperability of metadata between different collections may be facilitated. Slide 9 Dublin Core elements (attributes) Dublin Core includes 15 basic elements. When assigning metadata to resources, all elements are optional and repeatable. Some elements can be refined. For example, Abstract is a refinement of the Dublin Core Description element, making the element more specific. To support the interoperability goal, metadata authors must assume that refinements may be “dumbed down” if systems don’t support them. That is, an Abstract element may be mapped back to a Description element. Slide 10 Dublin Core values It is recommended that values for many Dublin come from controlled vocabularies. The CDP guidelines recommend specific vocabularies for different elements (such as the “DCMIType vocabulary” for the Type element). For example, the creator of metadata for a particular collection might specify that values for the Subject element must be selected from the Library of Congress Subject Headings (LCSH) and not by entering free keywords. Slide 11 Metadata for people, metadata for computers The “audience” for Dublin Core (or any other) metadata may be a person, or it may be a computer, using the metadata to facilitate search or do other processing. Slide 12 Dublin Core element descriptions The CDP guidelines provide an example of how Dublin Core elements have been interpreted for a particular set of institutions and their common project goals of providing access to digitized materials. You may find this format useful in thinking about how to create similar descriptions and guidelines for your schema assignment. Slide 13 Dublin Core Title element Label: Title Dublin Core definition: The name given to the resource. Dublin Core comment: Typically, a Title will be a name by which the resource is formally known. CDP comment: The name given by the creator or publisher; may also be an identifying phrase or name supplied by the contributing institution. Examples: Title=”Oral history interview with John Schulz" Title=”Untitled" Slide 14 Dublin Core Type element A cautionary tale Label: Resource Type Dublin Core Description: The nature or genre of the content of the resource. Dublin Core comment: Type includes terms describing general categories, functions, genres, or aggregation levels for content. Recommended best practice is to select a value from a controlled vocabulary (for example, the DCMIType vocabulary). To describe the physical or digital manifestation of the resource, use the Format element. CDP comment: Use only the DCMIType vocabulary. Examples: Type=”Moving image" Type="Sound" Type="Text" Type=”Collection” Slide 15 Summary • A schema is a set of attributes to describe a defined group of entities, along with associated value parameters and usage guidelines. We use the schema to produce metadata records that describe specific objects. • Dublin Core is a schema for describing information resources in a way that facilitates semantic interoperability between metadata systems. • Defining attributes in a way that makes it clear how to create records with them can be quite challenging, even for seemingly basic descriptive attributes. Slide 16 Some examples A data model for biomedical resources: Eagle-i Aggregations of cultural heritage data that incorporate Dublin Core in their data models: DPLA and Europeana Slide 17