Proposers: The RuleML initiative and the WSMO working group.
This Working Group is part of the Semantic Web Activity. It will focus on the development of a language to extend the semantic reach of current XML, RDF, and OWL meta-data efforts. In a recent talk on the Semantic Web, Tim Berners-Lee, Director of the W3C, has refined the layered structure for developing Semantic Web applications; in particular, the applications that make use of the rule-based representation. This working group will focus on building a rules layer, which will exist alongside with OWL, and will develop the formal underpinnings of such a language.
Rules constitute a key language layer, which is crucial to the success of the Semantic Web and its adoption by the industry. Rules are needed for policy specification, for specification of services and other complex entities on the Web, for querying, mediation, exchange of proofs, etc. A rules language that will adequately support such applications must go well beyond the expressivity of XML, RDF, and OWL, and it should build on the vast expertise and experience that was accummulated in the logic programming and database communities over the past thirty years. The rules language must have clear semantics, a natural user-friendly syntax, an XML-based interchange syntax, and RDF interoperability.
Specifically, the WebRules Working Group is chartered to design the following component:
Furthermore, the following general requirements must be met by the work produced by this Working Group:
The Working Group shall start by evaluating the technical solutions proposed in the RuleML, WRL, SWSL-Rules and SWRL drafts. If in this process the Working Group finds solutions that are agreed to be improvements over solutions suggested by these drafts, those improved solutions should be used.
The remainder of this section describes the requirements and deliverables in more detail.
The purpose of this WG is not to define a particular set of rules for use in any specific domain of discourse. Instead, the WG will define a machine-readable markup language, based on current Semantic Web standards, which will allow adopters to define their own rule sets. Such a standard language will facilitate the interoperability between tools and techniques that will be developed for manipulation of rule sets developed by the users of the language.
The term "rule" is interpreted slightly differently by different communities. In the context of this work, we refer to "standard" logic programming rules, i.e., rules that are based on Horn LP and are extended with default negation. Scoped default negation is a generalization of N3's log:notIncludes; it is similar to the usual default negation, but is explicit with respect to the scope that is used to close off the inference. (Such negation has been implemented in several systems, such as Flora-2 and Triple.)
Recently, a number of rules languages for the Web have been proposed. These proposals include SWRL, SWSL-Rules and WRL. All three proposals mostly use RuleML for the XML serialization of their languages.
Of these proposals, both SWSL-Rules and WRL are based on Horn LP; WRL additionally addresses the issues of integrity rules and interoperability with OWL, while SWSL-Rules addresses the issue of meta-programming. The Working Group will consider the built-ins proposed by SWRL and FORUM. The mentioned features of the languages should together be considered as a starting point for this particular WG product.
In addition to the actual specification of the language, the working group will deliver products that might be necessary for adoption of the specification. These will include use cases, reference implementations, manuals, and documents that describe the formal semantics of the language.
The Semantic Web community has several different visions of the nature of rules and their use. This working group will strive to be as general as possible in supporting multiple approaches and techniques. It should be noted, however, that the term "generality" is not used here in the same sense as "expressiveness," since the ability to say more things does not necessarily translate into the ability to support a wider community of users. The WG will have to consider trade-offs between expressivity, compactness, computational complexity, and simplicity.
More so than in case of a traditional programming language, a knowledge representation language needs a formal semantics to clearly delineate what is and is not entailed from any particular set of language constructs. Such a formal semantics should be based on a model theory. Evaluation strategies that are sound and (whenever possible) complete with respect to the semantics should be developed.
The new layered architecture for the Semantic Web places the rules language alongside OWL. SWRL, which is built on top of OWL, is considered complementary and out of scope for this working group. Other languages that are built on top of Description Logics, such as AL-Log, are also complementary and therefore out of scope.
To maximize the chances for adoption by the research and industrial communities -- the vendors and users of tools and products -- we need to focus on a core subset of features representing a must-have list of requirements for WebRules. We propose a phased approach to the development of the language, which will provide a scalable solution that could be augmented with additional features as needed. Our goal is to deliver a language of immediate use to the Web community.
In general, the Working Group will also consider items out of scope that are being addressed as part of other W3C activities, and it will reuse existing specifications wherever possible. The co-chairs and editors of this working group will coordinate with theit counterparts in the RDF Data Access Working Group (DAWG) to avoid duplicating work with respect to concrete data types, schema definitions, and other products which may fall into the areas where the Ontology (OWL) and rules come into contact. In addition, the working group chairs will work with the W3C's Semantic Web Coordination Group, particularly the Semantic Web Best Practices and Deployment (SWBPD) Working Group, to help keep our effort compatible with other ongoing Semantic Web development efforts.
The following are also considered out of scope for WebRules:
These are subject to revision due to editorial needs and external scheduling issues; updates will be negotiated with the related groups and recorded on the WebRules Working Group home page.
Last Call: September 2006
Proposed Recommendations: November 2006
Recommendation: January 2007
Errata / Wrap-up: May 2007
The group is expected to close down in mid 2007.
XML-derived activities, including the Semantic Web activities, have become strategic technologies for W3C and the IT industry at large. Each deliverable of any Working Group must satisfy the dependencies from other W3C Working Groups before it can advance to Candidate Recommendation.
At present, the primary relationship with other W3C work is through SparQL, RDF, and OWL activities. Coordination with these groups will be carried out on an ongoing basis within the W3C Semantic Web Coordination Group.
The Working Group will also attempt to liaise via cross-participation with a number of groups interested in rules and rules-based systems on the Web, as well as with work on Semantic Web Services, Web services and policy representation and interchange on the Web.
The WebRules Working Group will do its best to track the work done in as many other related groups as possible. Five groups have already been identified as doing related work:
To become a member of the WebRules Working Group, a representative of a W3C Member organization must be nominated by their Advisory Committee Representative (details on how to join are on the group home page). The nomination must include explicit agreement to this charter, including its goals, and the level of effort required and an IPR disclosure.
Membership is also open to invited experts from the community, selected by the chair or co-chair in order to balance the technical experience of the group.
Participation is expected to consume at least one day per week of each Working Group member's time.
The mailing list for group communication is www-webrules-wg@w3.org. An archive of www-webrules-wg is available to the public.
The Working Group will have a home page that records the history of the group, provides access to the archives, meeting minutes, updated schedule of deliverables, membership list, and relevant documents and resources. The page will be available to the public and will be maintained by one of the co-chairs in collaboration with the W3C team contact.
The Working Group will hold teleconferences approximately weekly. Participation in phone conferences is limited to members of the working group. The Chairs may, at their discretion, invite guest experts to attend particular phone conferences. An IRC channel may be used to supplement teleconferences.
Meeting records should be made available within two days of each telephone meeting.
Participation in face-to-face meetings is limited to working group members and observers invited by the Chairs. Observers may take part in decision-making at the discretion of the Chairs.
In addition to the required three annual face-to-face meetings, the Working Group may schedule other face-to-face meetings in a manner that maximizes co-location with events that Working Group members might be attending anyway.
The Chairs make Working Group meeting dates and locations available to the group at least eight weeks before the meeting, per W3C Process.
To be successful, we expect the Working Group to have approximately 10 to 20 active principal members for its 18-month duration. We also expect a large public review group that will participate in the mailing list discussions.
(W3C)
(W3C)