(Possible renaming: Defining Concepts and Relationships for the Semantic Web)
Publication of factual information on the Semantic Web often needs to be complemented by conditional information to communicate precisely what it means and/or how it may be used. Rules are an unambiguous and potentially machine-processable way of providing these extra conditions. Rules with variables also allow to generalize specific cases to (conditional) situation patterns.
A case where rules add clarity and precision to factual information is the publication of a new Semantic Web vocabulary, where the goal is to specify (proof-theoretic) semantics for some of the vocabulary elements with a low barrier to entry. Examples include the RDFS and SKOS specifications. Where appropriate, inference rules are expressed using the Jena 2 rule syntax, or as RDF statements using the OWL vocabulary. They are described in prose when this is not possible or when the expression would be too cumbersome to be a useful communication tool. A RIF provides implementation neutrality whereby the publisher can convey rule-based expression of semantics in implementation-neutral form, leaving the consumer free to choose whatever implementation approach is appropriate.
Rules may also be used to precisely communicate how data should be used. For example, a future FOAF user might publish not only unconditional information about their preferences but also conditional information, i.e. a rule set. This would complement their normal FOAF profile, e.g. describing which of their phone numbers should be used at which time of day depending on the day of the week:
Monday to Friday IF 09:00 - 12:00 OR 13:00 - 17:00, THEN call my office number IF 12:00 - 13:00 OR 17:00 - 18:00, THEN call my cell number IF 18:00 - 21:00, THEN call my home number IF 21:00 - 09:00, THEN do not call me Saturday, Sunday and Holidays IF 09:00 - 21:00, THEN call my cell number IF 21:00 - 09:00, THEN do not call me
In some cases, the publication of rules may more precisely communicate both what the accompanying information means and how it may be used. For instance, the IDM laboratory wants to publish an ontology of the anatomy of the brain cortex and a vocabulary to describe facts about items on brain images, such as can be provided by computer-assisted or automated image analysis tools. Hospital, medical practices and medical imagery laboratories want to use the ontology and vocabulary to improve and streamline their communication.
To communicate the semantics of the properties defined as part of the ontology and vocabulary in a precise and unambiguous way, IDM publishes a set of rules to capture relationships between ontology and/or vocabulary properties, e.g.:
Two MaterialAnatomicalEntities (MAE) having a shared boundary are connected: isMAEBoundedBy(?x1,?x3) Λ isMAEBoundedBy(?x2,?x3) Λ MAE(?x1) Λ MAE(?x2) Λ GyriConnection(?x3) -> isMAEConnectedTo(?x1,?x2) Two MAE entities having a shared connection are connected: connectsMAE(?x3,?x1,?x2) Λ MAE(?x1) Λ MAE(?x2) Λ GyriConnection(?x3) -> isMAEConnectedTo(?x1,?x2)
One MRI laboratory uses an automated image analysis tool to extract facts about brain images and wants to use the ontology and the vocabulary to describe items on the images. The rules are retrieved in the RIF and translated into the rule language specific to the engine used by the application. They are executed to automatically label the brain cortex structures - sulci and gyri - in the brain images, based on the facts extracted by the image analysis.