Rule Languages and Rule Interchange Sujit R Nair

Rule Languages and Rule Interchange Sujit R Nair November 30, 2009

Introduction �Need / Requirement. �Characteristics of current rule markup Languages. �A sample Scenario of Rule Interchange �Future Work. �Conclusion.

Need / Requirement �Expressiveness. �Machine-interpretation. �Automated Processing. �Translation and Interchange. �Publication and Serialization.

Current Rule Languages �Rule. ML. �SWRL. �R 2 ML. �W 3 C RIF. �Triple. �N 3. �Jena Rules. �Prova.

Rule. ML �Based on XML. �Rules consists of derivation, integrity constraints, production, reaction, tools and transformation. �Datalog Rule. ML and Hornlog Rule. ML.

SWRL �Combining Sublanguages of OWL and Rule. ML. �Union of DLP and Horn Logic. �Adds rules at the cost of undecidability and lack of complete implementation. �Protégé. �Rules = Antecedent + Consequent.

R 2 ML �Based on concept of Rule. ML and SWRL. �Rules consists of derivation, integrity constraints, production, reaction. �MOF/UML support.

W 3 C RIF �Facilitate rule exchange. �Dialects. �Basic Logic Dialect(RIF-BLD) ↔ Horn Rule Language. �Datatypes and Built-Ins(RIF-DTB). �Core. �Production Rule Dialect.

Jena Rules �Java Rule object. �Jena 2 is much simpler, allows rules to be specified in compact form in text source files.

Prova �Web rule language + Highly expressive distributed Web rule engine. �Supports complex reaction rule-based workflow, rulebased event processing, distributed inference services, rule interchange, rule-based decision logic, dynamic access to external data , Web Services and JAVA APIs.

Rule Interchange Sample � Using RIF to interchange F-Logic Rules and Drools. � Scenario : To supply a vendor-neutral representation of rules, so that rule-system developers can do their work without concern about a vendor-specific format and in particular without concern about the compatibility with the technology of its business partners. �Jane's e-commerce system uses Drools therefore the rules vocabulary is represented using Java beans �John uses an OWL vocabulary and its rules are represented in F-Logic.

Rule Interchange (contd) �The Two rules that are to be implemented : a. Rule R 1: If an item is perishable and it is delivered more than 10 days after the scheduled delivery date then the item will be rejected. b. Rule R 2: If an item is perishable and it is delivered more than 7 days after the scheduled delivery date but less than 14 days after the scheduled delivery date then a discount of 18. 7% will be applied to this delivery.

Rule Interchange (contd) �Points to be kept in mind: �RIF needs to deal with different vocabularies. �RIF needs an uniform mechanism to address vocabulary elements. The usage of URI's may be such a mechanism. �Vocabulary and rules are separate layers

Rule Interchange (contd) �John's OWL vocabulary: <rdf: RDF xmlns: rdf="http: //www. w 3. org/1999/02/22 -rdf-syntax-ns#" xmlns: rdfs="http: //www. w 3. org/2000/01/rdf-schema#" xmlns: owl="http: //www. w 3. org/2002/07/owl#" xmlns="http: //www. example. org/John. System/vocabulary"> <owl: Class rdf: ID="Item"/> <owl: Datatype. Property rdf: ID="perishable"> <rdfs: domain rdf: resource="#Item"/> <rdfs: range rdf: resource="xs: boolean"/> </owl: Datatype. Property>

Rule Interchange (contd) <owl: Datatype. Property rdf: ID="actual. Delivery. Date"> <rdfs: domain rdf: resource="#Item"/> <rdfs: range rdf: resource="xs: date"/> </owl: Datatype. Property> <owl: Datatype. Property rdf: ID="scheduled. Delivery. Date"> <rdfs: domain rdf: resource="#Item"/> <rdfs: range rdf: resource="xs: date"/> </owl: Datatype. Property> <owl: Datatype. Property rdf: ID="status"> <rdfs: domain rdf: resource="#Item"/> <rdfs: range rdf: resource="Status"/>

Rule Interchange (contd) </owl: Datatype. Property> <owl: Datatype. Property rdf: ID="discount"> <rdfs: domain rdf: resource="#Item"/> <rdfs: range rdf: resource="xs: double"/> </owl: Datatype. Property> <owl: Class rdf: ID="Delivery"/> <owl: Class rdf: ID="Status"> <owl: one. Of rdf: parse. Type="Collection"> <rdf: List> <rdf: first rdf: datatype="xs: string">rejected</rdf: first> <rdf: rest> <rdf: first rdf: datatype="xs: string">accepted</rdf: first> </rdf: rest> </rdf: List> </owl: one. Of> </owl: Class> </rdf: RDF>

Rule Interchange (contd) �John Rules (F-Logic) //If an item is perishable and it is delivered more than 10 days //after the scheduled delivery date then the item will be rejected RULE R 1 I: item[status -> "rejected"] <- I: item[is. Perishable -> true] AND I[actual. Delivery. Date > A] AND I[scheduled. Delivery. Date -> S] AND R is (A - S) AND R > 10.

Rule Interchange (contd) //If an item is perishable and it is delivered more than 7 days after the scheduled delivery date //but less than 14 days after the scheduled delivery date //then a discount of 18. 7% will be applied to this delivery. RULE R 2 I: item[discount -> 18. 7] <- D: delivery[has. Items>>Items] AND I: item[is. Perishable -> true] AND I[actual. Delivery. Date -> A] AND I[scheduled. Delivery. Date> S] AND member(I, Items) AND R is (A - S) AND R > 7.

Rule Interchange (contd) �Jane's Rules (Drools) package com. sample import com. sample. Item; rule "R 1" when i : Item(is. Perishable==true, actual. Delivery. Date : actual. Delivery. Date, scheduled. Delivery. Date : scheduled. Delivery. Date ) eval( actual. Delivery. Date. get. Day() - scheduled. Delivery. Date. get. Day() > 10) then i. is. Rejected(true); modify(i); end

Rule Interchange (contd) rule "R 2" when d: Delivery(has. Items: has. Items) i : Item(actual. Delivery. Date : actual. Delivery. Date, scheduled. Delivery. Date : scheduled. Delivery. Date ) eval( (actual. Delivery. Date. get. Day() - scheduled. Delivery. Date. get. Day() > 7) && has. Items. contains(i)) then i. set. Discount(18. 7); modify(i); end

Rule Interchange (contd) �Importing Jane's Rules in the John's rule system �Translating from Drools into RIF �Presentation Syntax proposed by RIF-BLD is used to perform the vocabulary interchange.

Rule Interchange (contd) �Translation of Jane's Drools Rules rules into RIF rule "R 1" when i : Item(is. Perishable==true, actual. Delivery. Date : actual. Delivery. Date, scheduled. Delivery. Date : scheduled. Delivery. Date ) eval( actual. Delivery. Date. get. Day() - scheduled. Delivery. Date. get. Day() > 10) then i. is. Rejected(true); modify(i); end

Rule Interchange (contd) // Translation to RIF Rule ("http: //jane. com/R 1" ? i#Item[(is. Rejected->"true"^^xs: boolean) : - And(? i#Item[(is. Perishable->"true"^^xs: boolean) (actual. Delivery. Date->? actual. Delivery. Date ) (scheduled. Delivery. Date ->? scheduled. Delivery. Date)] op: numeric-greater-than( op: numeric-subtract( userdefined: get. Day(? actual. Delivery. Date), userdefined: get. Day(? scheduled. Delivery. Date) ), "10"^^xs: int ) ) )

Future Work �On going research to develop a general rule mark up language. �Most of the research is concentrating on Rule. ML and RIF. �To cover the shortcomings such as – procedural call to external objects, operating systems , terminological descriptions for dealing with real world distributed Web Applications.

Conclusion �Why is rule markup needed ? �Currents rule markup languages. �Rule. ML and RIF. �W 3 C RIF BLD provides a robust core for the interchange of business rules among commercial and academic rule systems. � Embodies both production rules and derivation rules.

References Adrian Paschke , Harold Boley – Rule Markup Languages and SWRL. 2. W 3 C Rule Interchange Format, http: //www. w 3. org/2005/rules/wiki/RIF_Working_Grou p 3. RIF Basic Logic Dialect, http: //www. w 3. org/TR/rif-bld/ 4. W 3 C RIF Use Cases and Requirements, http: //www. w 3. org/TR/rif-ucr/ �Gary Hallmark , Christian de Sainte Marie, Marcos Didonet Del Fabro, Patrick Albert , and Adrian Paschke - Please Pass the Rules: A Rule Interchange Demonstration 1.

Contact Information Sujit Raveendran Nair. Email: srn 071000@utdallas. edu. Phone: 972 -983 -4886.