Metadata Basics

Making a statement

Making a statement

Most metadata today is gathered into machine-readable records. Each record represents a thing being described. The fields that make up the description relate to the subject of the record. Oftentimes, the subject of the record is to be interpreted as "the thing that all of these fields describe." There is no explicit identity for the subject of the description other than, perhaps, the record's identifier.

a record

The record acts as a container for the descriptive fields. This has the advantage that it keeps the fields together for distribution and use. But it also has the disadvantage that the data in the fields cannot be used except within the context of that record. This limits the interoperability of data, especially where different systems use different record formats. In other words, you can't take

color = red

and do anything with it, even though there may be many other bits of data in the world that make a statement about "color" because outside of the context of the record you don't know what this statement refers to. It is essentially like a sentence without a subject: "? is red."

To make your metadata interoperable you need to break the record apart and give each metadata statement a subject.


This is the same information as in the record, but now each individual statement has a subject, a predicate, and an object. The subject is always a thing that has a unique identifier. So let's create an identifier for our subject:

We can now say

[] | color | red
[] | make | Chevrolet

etc. Unlike the use of an identifier to identify a record (that is, a particular set of metadata), the identifier in this case identifies the subject of the metadata. If we want to add more metadata about this subject, we create another statement with the same subject identifier.

Each individual statement is now complete, and any statement can be used in combination with other statements about the same subject, and can even be combined with data from other sources that describe the same subject. At the same time, we can also gather together all of the information about this subject, even if it comes from different sources by selecting all statements with the same subject. This statement structure is the basis for the Semantic Web. It allows you to easily mix metadata from different sources and to create webs of data.