| SubmissionsW3C is pleased to receive the OWL-S Submission from:
This Submission proposes a framework based on the OWL Web Ontology Language to help users and agents search, discover, invoke, compose and monitor Web services. It includes 8 ontologies written in OWL as an extensible core.
This is the first Submission in the area of Semantic Web Services, where Semantic Web technologies are applied to the challenges offered in the Web services arena. Details of the OWL-S work have been extensively discussed in the W3C Semantic Web Services Interest Group and the Semantic Web Services Initiative.
In OWL-S, information used for Web service automation is exchanged in the same language the rest of the web is beginning to use for expressing more and more data: RDF. We look forward to new functionality and user benefits deriving from this emerging interoperability of data.
The name "OWL-S" and the title OWL Web Ontology Language for Services may be confusing to some readers. The OWL-S Submission specifies a framework of ontologies written in OWL, not an ontology language or a version of OWL. The current name evolved out of the historical structure of the DAML program, and we suggest that any future work proceed under a name more precisely descriptive of the functionality provided.
The OWL-S Submission groups information about a service into three main ontologies:
The Profile is meant to be used by publishing and discovery applications, providing search and matchmaking. It includes expression of limitations, requirements, and quality of service.
The Process describes the use of the service, detailing the semantic content of messages. This semantic description aims at helping composition, coordination, and execution monitoring of Web services.
The Grounding is the link to the service's concrete realization. It defines a mapping of "atomic processes" in OWL-S to WSDL input and output messages. This binding from the abstract, semantically-described processes to concrete exchanges specification is essential to begin integration with the existing Web services architecture.
The Service ontology helps organize the above parts, while the other four ontologies in the submission serve in supplementary roles.
New points of integration with Web services specifications under development are likely to require additional or updated ontologies.
The Submission includes a document detailing OWL-S' Relationship to Selected Other Technologies.
The OWL-S Submission has a strong dependency on the Web Services Description Language (WSDL) 1.1 Submission. Several parts of this framework rely on mappings to WSDL: an "atomic process" in OWL-S corresponds to a WSDL 1.1 operation, sets of inputs/outputs used in such an atomic process map to WSDL messages, and types map to WSDL abstract types. The Grounding ontology achieves the link to WSDL. Additional coordinated work would be needed to transition OWL-S to use WSDL 2.0, which is currently under development.
The Profile ontology overlaps the area of constraints and capabilities, as discussed at the recent W3C workshop. Expressing conditions of use and applicability, as well as quality of service, represent a significant part of this issue.
The design of OWL-S also brings to Web services choreography and orchestration an approach to automated service composition, coordination, and cooperation based on description logic (OWL-DL). This work is related to the Web Services Choreography Description Language (WS-CDL).
The design of OWL-S poses some opportunities and challenges as a major early user of OWL and a potential user of a rule language. OWL-S currently uses Semantic Web Rule Language (SWRL) 0.6 in an experimental capacity. The Web services user base and use cases may help focus rule language activities within the W3C, which has staff involved in related research (cwm, metalog), supports mailing lists (www-rdf-rules, www-rdf-logic), and is in the early stages of planning a Semantic Web rule language workshop to be held in the first half of 2005.
We look forward to the Web services community providing feedback on the utility of the architecture and overall functionality offered by OWL-S. We also look forward to a discussion of the degree of interest that exists in integrating the OWL-S approach into the mainstream of Web services technologies, as well as the strategy and sequencing that could be put in place to accomplish this goal.
Discussion is welcome within the Semantic Web Services Interest Group about OWL-S along with other approaches to Semantic Web Services, such as Web Services Modeling Ontology (WSMO) and related strategic issues. We intend to hold a Semantic Web Services workshop in the first half of 2005. One possible outcome would be a Recommendation track on a Semantic Web Services framework.