A metadata schema is a list of elements or defined data points that are used to capture information about a resource. Some of these data points might include a title, an identifier, a creator name, or a date.
Standards are the tools that tell us how to populate each of the data elements within a schema. There are three types of standards:
Content standards describe the use of each element or what pieces of types of information go where. They also give guidance on how to best record or transcribe that information: Where should that information be coming from? What is the best source of information? Which elements should require the use of data values standards and, if so, which value standards should be used?
Data value standards are lists of standardized subject terms, genre terms, names, etc. Examples of data value standards include controlled vocabularies like Library of Congress Subject Headings and other discipline-specific thesauri, and encoding or formatting standards like ISO 8601 which prescribes how dates ought to be represented.
Data structure standards (XML, RDF, etc.) exist to tell us how to encode and structure the metadata record so as to ensure its machine readability. It is important to note that there are primary two users of metadata: humans and machines. Any metadata you create must be intelligible to both.
Best practices guide or prescribe the selection and use of a particular metadata schema, which subset of elements within that schema will be used to comprise a metadata profile for use in describing a certain class or collection of objects. Best practices also serve to: 1) better define or clarify within the context of a given project or organization the use of each metadata element; and 2) more decisively recommend the use of particular content, data value, and data structure standards.
When creating or evaluating metadata it is important to ask: