From W3C Wiki
Jump to: navigation, search

SW-UbiComp-SIG is an open community for exploring the use of Semantic Web technology in building Pervasive/Ubiquitous Computing applications.

The group's current goal is to define a standard ontology to support pervasive computing applications (see our web site for the draft proposal). This includes OWL ontologies for person, device, time, space, event, digital documents, and policy (for security and privacy).

SW-UbiComp-SIG welcomes everyone who is interested in the development of Semantic Web and Ubiquitous Computing to participate in discussions and make contributions.

This wiki page serves as a "scratch pad" for the community to share ideas. Feel free to add/modify any contents on this page. We also have a mailing list on the Yahoo Groups.


Q: How can i reuse the ontologies designed/developed here with the ontology toolkit Protege from Stanford?

Q: What's goal of the PERVASIVE-SO project?

A: Our goal is to define useful ontologies for supporting pervasive computing application. Vocabularies will be included in the ontology if they are useful for demonstrating some aspect of the pervasive computing vision. We are less focused on building a comprehensive ontology library for pervasive computing.

Q: What vocabularies should be covered in the Pervasive Computing Standard Ontology (PERVASIVE-SO)?

A: At present we plan to develop ontologies for agent, person, device, time, space, event, document, and policy.

Q: What design strategy should we adopt when developing PERVASIVE-SO?

A: We will define use cases to motivate our ontology design. In each use case, we describe some pervasive computing application that can be implemented to enhance our computing life style. Based on the functional requirements of these applications, we identify ontology vocabularies that are needed to support the implementation. Identified key vocabularies will be included in the PERVASIVE-SO ontology.

HarryChen: Simplicity is the key. We should focus on building ontologies that are simple, easy to understand and use. While sometimes it is necessary to introduce abstract concepts, which often adds additional layers of complexity, our goal is to keep the number of these concepts minimal. Don't add new concepts just because they can be easily written down. Without knowing how these concepts can be useful, adding them to the ontology will just make things difficult.

In short, Keep It Simple, Stupid & Less is More. Most importantly, DontWorryBeCrappy.

Q: What's the merit in defining vocabularies that overlap with other ontologies (e.g., DAML-Time, DAML-Space, FOAF)?

A: At present, one of the greatest challenges is to build systems that actually reuse multiple independently developed ontologies. Because it's relatively easy to prototype systems that only use ontologies that come from a single source, the PERVASIVE-SO project is about building a common ontology by combining the vocabularies from different consensus ontologies.

HarryChen: A great challenge in today's Semantic Web world is ontology reuse -- reuse ontologies that are developed by other people. Often we find ourselves "re-inventing" ontologies that have already been developed by other people, but yet we continue to convince ourselves that we should develop our own ontologies. I think there are at least two reasons why we do this: (i) It's more fun to write ontologies than using ontologies (we lack tools for building large scale ontology-oriented systems). After writing our own ontologies, we feel that we have accomplished something. (ii) Many of the existing ontologies lack proper maintenance. Often people fail to check for the validity of their ontologies before publishing them on the web and fail to develop proper version control schemes for publishing their ontologies. Given these two reasons, I believe in order bring Semantic Web ontologies to the next level, we should practice (or learn how to practice) better ontology document maintenance.

Use Cases

The following use cases describes a set of pervasive computing applications that can be implemented to enhance our computing life style. The purpose for defining these use cases is to motivate the PERVASIVE-SO development.

For the use case authors: In addition to describing the use case, you should also include a description of the ontology that you anticipate to be needed to support the use case. This description may just be a list of ontology types (e.g., person ontology, device ontology).

Suggestions and Comments

Comments on building PERVASIVE-SO ontologies

  • Use bnode to represent a per:Person; how about demonstrating the use OWL ontology mapping constructs (see Danny Ayers's message)