Topic maps are abstract structures that can encode knowledge and connect this encoded knowledge to relevant information resources. Topic maps are organized around topics, which represent subjects of discourse; associations, representing relationships between the subjects; and occurrences, which connect the subjects to pertinent information resources.

Topic maps may be represented in many ways: using topic map syntaxes in files, inside databases, as internal data structures in running programs, and even mentally in the minds of humans. All these forms are different ways of representing the same abstract structure, the Topic Maps Data Model defined in Part 2 of this standard.

Canonicalization is the process of serializing a data structure in such a way that two data structures considered to be the same result in the same serialization and two data structures not considered to be the same result in two different serializations. A canonical form enables direct comparison of two data model instances to determine equality by comparison of their canonical serialization.