This is a description of one of the general General Use Case Categories abstracted from the possible Use Cases.
1. Abstract
This category concerns the use of RIF for publishing how a semantic web vocabulary or set of instance data should be interpreted.
The essence of 'publication' is that the publisher is using the rules as a vehicle for communicating some aspect of what they mean by the accompanying semantic web information or how it may be used. The end consumer may not necessarily use a rule-based system for implementing or complying with the published semantics. The publication is somehow an end in itself.
In the case of a vocabulary the role of rules is to define a proof-theoretic semantics for the vocabulary. Examples of this include the RDFS and SKOS specifications. Rules are a relevant tool for this where the given semantics is not expressible in OWL or where the expression would be too cumbersome to be a useful communication tool.
In the case of instance data the role of rules is to publish how that data should be used or interpreted. For example a FOAF user might publish a rule set along with their FOAF record describing which of their phone numbers should be used at which time.
2. Status
Based on Use Cases originally proposed by DaveReynolds and by HaroldBoley each of which cite existing usage, supporting rule systems and ongoing implementations.
Narrative based on Use Case proposed at workshop by Christine Golbraich added by ChristianDeSainteMarie.
3. Links to Related Use Cases
Many other use cases involve publication of rules expressing policy Policy-Based Transaction Authorization and Access Control or usage constraints Regulation/Constraint Compliance Monitoring, perhaps the feature that distinguishes this category is that publication is an end in itself (communication of the rules)
The second use case (FOAF RDF Rules) also has aspects of defining derivations that can be made for a FOAF record and so is also related to Rich Knowledge Representation
4. Relationship to OWL/RDF Compatibility
These cases all concern publication of rules to augment RDF or OWL data or vocabularies and so require RDF and OWL compatibility.
5. Examples of Rule Platforms Supporting this Use Case
CWM and JenaRules have been used to express vocabulary semantics for SKOS
- OO jDREW is being used for implementation of the FOAF rules use case
6. Benefits of Interchange
- Implementation neutrality - publisher can convey rule-based expression of semantics in implementation-neutral form leaving the consumer free to chose whatever implementation approach is appropriate.
- Rules published may be shared, harvested and reused by other publishers (e.g. other individuals in the FOAF case).
7. Requirements on the RIF
- Compatibility with base semantic web languages - RDF, OWL.
- A human (or at least developer) legible syntax (in addition to the machine processable syntax).
- For publication in associated with an RDF data set (e.g. FOAF) a closed world assumption scoped to that data set may be required.
- Typically the data set or new vocabulary will draw on other (OWL) ontologies and so the rules need to be able to reference and exploit those ontologies.
8. Breakdown
8.1. Actors and their Goals
- Publisher - publishing a semantic web data set or vocabulary they wish to augment with rules (ranges from standards body to individual provider)
- Consumer - wishes to use the vocabulary or data set (ranges from tool implementer to individual data consumer)
- Aggregator - optional third party aggregator that mediates the interaction between the publisher and consumer, wishes to convey the integrated data + rule set to the consumer, also wishes to facilitate other parties discovering and reusing interesting rules.
8.2. Main Sequence
- Publisher creates data set or new vocabulary definition and associated set of rules defining usage or interpretation.
- Publisher makes rules set available (on web, via web service ...) along with base data set.
- Consumer finds data set and rules, perhaps with assistance of third party aggregator.
- Consumer applies rules (this last step is rather different in different specific cases)
9. Narratives
Describe possible scenarios illustrating the use case in separate subsections, assigning a title to each.
9.1. Labelling of brain anatomical structures
This narrative is taken from the contribution presented by Christine Golbreich et al. at the W3C Workshop on Rule Languages for Interoperability .
The IDM laboratory publishes an ontology of the anatomy of the brain cortex (currently as an HTML document, to be migrated to OWL-DL soon). In addition to the ontology classes that represent the brain anatomical entities organized in a hierarchy, the ontology defines mereological as well as topological properties. Rules are used to capture ralationships between ontology properties, e.g.:
Two MaterialAnatomicalEntities (MAE) enving a shared boundary are connected: isMAEBoundedBy(?x1,?x3) Λ isMAEBoundedBy(?x2,?x3) Λ MAE(?x1) Λ MAE(?x2) Λ GyriConnection(?x3) -> isMAEConnectedTo(?x1,?x2)
One of the purported usage of the ontology is to help in the labelling of the brain cortex structures - sulci and gyri - in brain images (e.g. MRI). IDM defined n-ary domain relations, such as the ternary predicates connects and separates or the binary predicate hasNoCommonPart. These relations do not belong to the ontology, but are useful to describe facts about items on brain images. Rules are used to capture how these so-called domain properties relate to ontology properties and how they can be propagated, e.g.
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) A sulcus having a segment separating two material entities separates them too: separatesMAE(?y, ?x2, ?x3) Λ hasSegment(?x1,?y) Λ Sulcus(?x1) Λ MAE(?x2) Λ MAE(?x3) Λ SF(?y) -> separatesMAE(?x1, ?x2, ?x3)
On the one hand, the rules are used to specify semantics for some of the elements of the published ontology in a clear, precise and unambiguous way. On the other hand, they are used to specify how factual information about elements in brain images, such as can be provided by automated image analysis tools, can be used to identify and label anatomical entities in the images.
The benefit of publishing the ontology in a machine-processible form - e.g. for use by medical imagery laboratories and medical pratices - thus depends greatly on the availability of a standard RIF to publish the rules alongside it.
9.2. FOAF publication of usage rules
Ben has created his person-centric metadata in the form of a FOAF document as has become popular, covering unconditional properties about his personal interests, friends, etc. However, he also needs to specify properties conditional on other persons, the time, the location, etc. For example, Ben prefers to specify his work time and free time s.t. his colleagues can contact him when at work, while his friends can reach him when off work. Consequently, with rules, his schedule might look as follows:
Monday - Friday
- IF 09:00 AM - 12:00 AM, THEN call my office number IF 13:00 AM - 17:00 AM, THEN call my office number IF 18:00 AM - 21:00 AM, THEN call my home number IF 12:00 AM - 13:00 AM, THEN call my cell number IF 17:00 AM - 18:00 AM, THEN call my cell number IF 21:00 AM - 09:00 AM, THEN do not call me
Saturday, Sunday, Holidays
- IF 09:00 AM - 21:00 AM, THEN call my cell number IF 21:00 AM - 09:00 AM, THEN do not call me
10. Commentary
Of the two use cases citied in this group, this 'publication' abstraction does not fully cover the FOAF rules case, which also involves execution of those rules - see related use case groups.
One the the most talked about cases of mass publication of rules is probably privacy policies, but that's presumably covered under Policy-Based Transaction Authorization and Access Control.
You can also imagine people publishing rules on their own, not associated with data or ontologies, just for the purposes of reuse. Analogous to "here's a useful XSLT script I cooked up".