SmartContacts

Yong Liu, Ph.D. Candidate in Computer Science

Kay Connelly, Assistant Professor in Computer Science

The number of cell phone users has doubled since 2000 [1]. However, because of the static nature of their notification mechanisms, cell phones often create interruptions to receivers such as ringing in a meeting or movie theatre. Previous research shows that such interruptions may lead to an increased level of receivers’ stress and errors [2] and cause annoyance to people in the surrounding environment [3]. Various approaches have been proposed to solve the problem of inappropriate incoming calls. Among them is a category known as caller-based solutions [4, 5], which provides the caller with the context of the receiver. Examples of context include a person’s location, the activity in which they are engaged, or the number of surrounding people.

The caller-based approach to minimizing inappropriate cell phone interruptions assumes that communication is a cooperative act, and leverages the caller’s judgment in determining whether it is an appropriate time to call. This type of solution requires the receiver’s social context either to be sensed automatically, or manually set by the receiver, and then delivered to the caller. Current proposed solutions do not address the infrastructure required for the caller’s mobile device to automatically obtain the receiver’s context information without human intervention and assume some type of manual bootstrapping. Our approach focuses on this automatic discovery using the existing Web infrastructure.

We are developing a large scale social context service discovery system called SmartContacts. Drawing on information retrieval technology and the established Internet infrastructure, SmartContacts provides a platform for cell phone users to publish their own context and retrieve other users’ context, if allowed by the information owners. It has global automatic context discovery and utilization capability.

SmartContacts Approach

SmartContacts works in a way similar to retrieving a piece of information from billions of web pages on the Internet. SmartContacts enables cell phone owners to publish their context in the form of a web service and a service description document on a web server of their choice. The description document is then crawled and indexed by major Internet search engines like Google. When a caller would like to know the context of a specific receiver, his SmartContacts client finds the address of the corresponding web service via Internet search engines and retrieves the context through the service handle. The entire discovery process is automatic. All the caller needs to know is the phone number of the receiver.

The service discovery mechanism adopted in SmartContacts offers the following advantages:

• High scalability – services are organized in a decentralized way
• Wide area service discovery – services are discovered Internet wide by the existing Internet infrastructure
• Automatic service discovery and consumption – services are discovered and consumed according to their descriptions, e.g., a cell phone number
• High fault-tolerance – there is no single point of failure in the discovery system maintenance

• Context Publisher is a program on the receiver’s cell phone which automatically publishes the current context.
• SmartContacts Client is a program on the caller’s cell phone that retrieves a receiver’s current context. It uses the receiver’s cell phone number to retrieve the information on demand in real time without human intervention.
• Web Server hosts context Web services for one or more cell phone users. It can be any third party Web server and does not need to be dedicated for SmartContacts.
• Web Search Engine is an existing search engine which is used to locate the service description document of a specific phone, from which a service handle may be attained.

Service/context publishing:

• Registration: Users publish their context querying Web service WSDL file and a description HTML document containing the address of the WSDL file to a third party Web server.
• Discovery: The description HTML document is crawled and indexed by major Internet search engines.
• Publication: The Context Publisher runs as a background application, periodically checking the context information and updating the third party Web server via the corresponding Web service.

• Search Engine Query: SmartContacts client issues a query to an arbitrary Internet search engine via some interface provided by the search engine. The keyword used in the query contains some information which uniquely identifies the receiver, such as cell phone number.
• Context Web Service Discovery: Once the corresponding HTML document is obtained, the SmartContacts client parses the HTML to get the address of the Web service WSDL file and download it.
• Context Query: By analyzing the WSDL file, the client knows where and how to invoke a remote Web service method to retrieve the context information.

We have developed a SmartContacts prototype on the Microsoft Smartphone platform. We use Google as the third party Internet search engine and the Google Web service API to obtain other SmartContacts users’ context Web service handles. SmartContacts users need to apply for their own Google API license key before starting to issue their queries via the SmartContacts client.

Currently, the SmartContacts Client interface is a separate application. We intend to integrate it with the current address book, so that the context can be displayed in the address book, and not using a separate interface.

In our current SmartContacts implementation, the context Web service is found by searching on the phone number. To provide the receiver with some privacy protection, a more complicated keyword should be used, such as a string with both the receiver’s phone number and name encoded in it. This can avoid context information from being exposed to strangers who search cell phone numbers randomly via a search engine. Further protection could be provided with the receiver providing some type of capability to callers, but this would require an offline capability exchange.

[1] REUTERS, Mobile phone users double since 2000. Dec, 2004.

[2] Eyrolle, H. and J. Cellier, The effects of interruptions in work activity: Field and laboratory results. Applied Ergonomics, 2000. 31: p. 537-543.

[3] Wei, R. and Leung, L., Blurring public and private behaviors in public space: policy challenges in the use and improper use of the cell phone. Telematics and Informatics, 16, 1999, 11– 26.

[4] Tang, J.C., et al. ConNexus to awarenex: extending awareness to mobile users. In Proceedings of CHI 2001.

[5] Schmidt, A., Takaluoma, A., and Mantyjarvi, J., Context-Aware Telephony Over WAP. Personal and Ubiquitous Computing, 4(4), 2000, 225-229.


	People Yong Liu, Ph.D. Candidate in Computer Science Kay Connelly, Assistant Professor in Computer Science


	Introduction The number of cell phone users has doubled since 2000 [1]. However, because of the static nature of their notification mechanisms, cell phones often create interruptions to receivers such as ringing in a meeting or movie theatre. Previous research shows that such interruptions may lead to an increased level of receivers’ stress and errors [2] and cause annoyance to people in the surrounding environment [3]. Various approaches have been proposed to solve the problem of inappropriate incoming calls. Among them is a category known as caller-based solutions [4, 5], which provides the caller with the context of the receiver. Examples of context include a person’s location, the activity in which they are engaged, or the number of surrounding people. The caller-based approach to minimizing inappropriate cell phone interruptions assumes that communication is a cooperative act, and leverages the caller’s judgment in determining whether it is an appropriate time to call. This type of solution requires the receiver’s social context either to be sensed automatically, or manually set by the receiver, and then delivered to the caller. Current proposed solutions do not address the infrastructure required for the caller’s mobile device to automatically obtain the receiver’s context information without human intervention and assume some type of manual bootstrapping. Our approach focuses on this automatic discovery using the existing Web infrastructure. We are developing a large scale social context service discovery system called SmartContacts. Drawing on information retrieval technology and the established Internet infrastructure, SmartContacts provides a platform for cell phone users to publish their own context and retrieve other users’ context, if allowed by the information owners. It has global automatic context discovery and utilization capability.


	SmartContacts Approach SmartContacts works in a way similar to retrieving a piece of information from billions of web pages on the Internet. SmartContacts enables cell phone owners to publish their context in the form of a web service and a service description document on a web server of their choice. The description document is then crawled and indexed by major Internet search engines like Google. When a caller would like to know the context of a specific receiver, his SmartContacts client finds the address of the corresponding web service via Internet search engines and retrieves the context through the service handle. The entire discovery process is automatic. All the caller needs to know is the phone number of the receiver. The service discovery mechanism adopted in SmartContacts offers the following advantages: • High scalability – services are organized in a decentralized way • Wide area service discovery – services are discovered Internet wide by the existing Internet infrastructure • Automatic service discovery and consumption – services are discovered and consumed according to their descriptions, e.g., a cell phone number • High fault-tolerance – there is no single point of failure in the discovery system maintenance


	Current Progress We have developed a SmartContacts prototype on the Microsoft Smartphone platform. We use Google as the third party Internet search engine and the Google Web service API to obtain other SmartContacts users’ context Web service handles. SmartContacts users need to apply for their own Google API license key before starting to issue their queries via the SmartContacts client. Currently, the SmartContacts Client interface is a separate application. We intend to integrate it with the current address book, so that the context can be displayed in the address book, and not using a separate interface. In our current SmartContacts implementation, the context Web service is found by searching on the phone number. To provide the receiver with some privacy protection, a more complicated keyword should be used, such as a string with both the receiver’s phone number and name encoded in it. This can avoid context information from being exposed to strangers who search cell phone numbers randomly via a search engine. Further protection could be provided with the receiver providing some type of capability to callers, but this would require an offline capability exchange.


	References [1] REUTERS, Mobile phone users double since 2000. Dec, 2004. [2] Eyrolle, H. and J. Cellier, The effects of interruptions in work activity: Field and laboratory results. Applied Ergonomics, 2000. 31: p. 537-543. [3] Wei, R. and Leung, L., Blurring public and private behaviors in public space: policy challenges in the use and improper use of the cell phone. Telematics and Informatics, 16, 1999, 11– 26. [4] Tang, J.C., et al. ConNexus to awarenex: extending awareness to mobile users. In Proceedings of CHI 2001. [5] Schmidt, A., Takaluoma, A., and Mantyjarvi, J., Context-Aware Telephony Over WAP. Personal and Ubiquitous Computing, 4(4), 2000, 225-229.