Multilingual Linked Open Data Patterns

Label Everything

Description

Linked data datasets should provide labels for all resources: individuals, concepts and properties, not just the main entities.

Context

Linked data applications that contain data which is supposed to be exposed to an end-user in some natural language.

Although URIs may be human-readable, they are not expected to be seen by the end user. In order to improve user experience it is necessary to expose data and entities in human-readable ways.

Labels facilitate:

Discussion

In general, associating labels is a good idea. It is always better to offer a textual information to the end user than a URI. However, in some applications it can be difficult to find the right label for a resource, specially when resources are automatically generated.

When labeling resources, one must take into account that the purpose of those labels is mainly for humans. Using camel-case or similar notations should be avoided and the use of uppercase, space delimiters etc. should be consistent [Montiel 11].

See also

[Ell 11] enumerates some uses for labels and contains a thorough study on the use of labels in popular datasets.

This pattern also appears with the same name in the linked data patterns book, where dataset creators are urged to Ensure that every resource in a dataset has an rdfs:label property.