ICEBoxHome Networking

The ICEbox (Installation, Configuration, and Evolution box) is a network appliance that acts as a logical front door to the home network, serving as a central point of control for the home network, and providing a unified interface for easy home network configuration and management that shilelds the user from the technical details of the network. The ICEbox addresses common problems with device configuration, network security, and monitoring and troubleshooting. A simple physical pointing interface is used for initial client device configuration; a graphical interface combined with physical controls on the ICEbox provides access to management functions, including network security. Results of a series of our user studies indicate that our study participants found this appliance both useful and usable as a network configuration and management tool.

Home Media SharingHome Networking

Consumer digital media technologies and the Internet have brought a vast and diverse range of new technologies for media sharing. These technologies include not only simple person-to-person email sharing, but also a plethora of web-based systems designed to allow people to upload photos that can then be viewed by family, friends, or even complete strangers. However, previous research by our group (some of which is reported in Miller and Edwards (2003)) has demonstrated that many of these technologies do not adequately meet the needs of large numbers of consumer photographers. In this project, we explore a new home-to-home media sharing architecture which provides a platform for usable media sharing and experimentation with new modalities of sharing for the "Kodak Culture" people. Through our architecture, we try to address a couple of the needs of the "Kodak Culture" people inferred from empirical studies, including natural naming for media sender and receiver, implicit security and privacy, presentation and interactiveness, and so on.

Smart Home Network RouterHome Networking

We have created a "smart home router" device using an existing low-cost commercial router (Linksys WRT-54G), which aims to provide a suite of functionality specifically designed to go beyond basic routing, to address common network problem areas for users, thereby reducing support calls to vendors and increasing their ability to differentiate their products from competitors. Currently we have been adding three features:

1. Zero configuration anti-phishing support: our router ensures the anonymity of the user and allow for direct filtering of HTTP requests.
2. Network troubleshooting: our router analizez network traffic and the conditions of network components, and to use heuristics to provide common troubleshooting advice to users. For example, software on the router can determine if a problem lies with the user's ISP (loss of connectivity to the ISP, for instance, of failure of the ISP's domain name services) versus with a device on the home network. Users on the home network can simply pull up a web page to get real-time diagnostics and advice, and without having to resort to a customer help line, or the painful process of manual troubleshooting.
3. Network product advisor: our router implements a set of heuristics that can offer specific upgrade advice based on the specific characteristics of the user's actual network. For example, if the system detects a user trying to stream high-quality video over a wireless network, it may suggest an upgraded wireless technology that can better accommodate the user's needs; poor observed signal strength to client devices may prompt a suggestion that the user needs to add a wireless repeater. Other sorts of suggestions are possible. This mechanism provides a way for users to get specific suggestions based on their actual use of the network.

Speculative Designs for Home Networking EquipmentHome Networking

Home networking equipment is often unsightly. Hence householders frequently hide these devices in closets, behind books, or under sofas. We have created a set of speculative designs for home network routers that encourage people to "get their routers out from under the couch" and into visible spaces of the home. In addition to just making the hardware more aesthetically pleasing, these designs also incorporate ambient information displays intended to provide householders with a general sense of the health of their network, as well as basic help in diagnosing and troubleshooting network problems.

Empirical Studies of Home NetworkingHome Networking

We are seeking to understand the causes of digital complexity in the home and how householders cope with these difficulties. Our work includes in-home qualitative investigations, analyses of householder-generated sketches of their home networks, and analyses of phone calls to professional technical support hotlines.

Support for Collaborative Network TroubleshootingHome Networking

Despite rapid adoption of networked computing in the home, home networking setup and maintenance tasks are extremely difficult for many people. Recently, the Human-Computer Interaction research community has had an emerging interest in studying the user experience (e.g. usability problems) associated with networked computing and AV equipment in the home. However, this community has little empirical data about the help-seeking practices of people who are experiencing trouble setting up or fixing their networked equipment. We are collecting data about help-seeking by (1) automatically analyzing the content and context of popular online help forums and (2) conducting a set of focus groups about help-seeking practices. Through this research, we are discovering which aspects of home network setup and maintenance householders find most troubling, as well as how they seek help to resolve these problems.

Home Network Visualization ToolsHome Networking

Building and maintaining a home network is difficult. One of the key reasons for this difficulty is the invisibility of many salient features of the network, including its structure, state, and the configuration of devices on it. We believe that bringing visibility to the intangible, invisible aspects of home networking can empower householders to better understand and manage their home networks. We have prototyped software that has allowed us to explore the interplay between infrastructure visibility and home network troubleshooting.

Sesame: Informing Security Decisions with VisualizationUsable Security

Sesame brings a direct manipulation graphical interface to end-user security to help non-experts make better informed security choices. Most prior work in end-user security tends to be highly task or threat-specific; or largely text-based and/or designed for experts. There are few tools if any which focus explicitly on helping non-expert users to better understand the technical, system-level concepts needed in making security decisions. With Sesame, we seek to provide non-experts with a general, firewall-like tool that addresses a wide range of threats independent of specific tasks or applications.

Firefly: Towards a Visual Firewall Tool for End-UsersUsable Security

The personal firewall, like anti-virus software, has become an essential security tool for end-users; but unlike anti-virus tools, using a personal firewall requires end-user involvement. That is, its use cannot be entirely automated away through means such as automatic signature updates. With the Firefly project, we first seek to define more clearly what user involvement means with regard to firewall usage. Some questions we hope to answer include: Are there firewall mechanisms we can automate further to avoid burdening the user? What firewall decisions explicitly require input from the user? After better defining the user's decision domain, we then seek to design a visual firewall that will help to inform these decisions. The visual firewall interface design will build on prior work completed in the Sesame project.

Using Social Networks for Computer FirewallsUsable Security

End-user tools for network security management are infamous for their lack of usability. Common firewalls, for instance, constantly interrupt users with requests to allow or deny connections to certain IP addresses and port numbers; users rarely know the "correct" answers to these questions, nor even what they mean in many cases. Likewise, techniques such as human comparison of hash codes to verify program integrity are difficult to use. To address these problem, the Bonfire Project is developing a new Internet firewall system that leverages collaborative filtering techniques, much like Flickr or del.icio.us bookmarks, coupled with a unique hash code visualization technique. When confronted with a network security decision, users can at-a-glance determine what others have done in similar situations, as well as look at "tags" others have used to mark particular programs; easy-to-describe hash code visualizations allow them to verify program integrity easily. This tool brings into the interface the formerly informal network of acquaintances that users often enlist (through phone calls or email) when making decisions about system configuration.

TALC: Using Graffiti to Fight Software VulnerabilitiesUsable Security

Users are often unaware of the threats they face from constant broadband connectivity. These threats -spyware, malware, and so forth -can arise from a number of system configuration details: is the users' software fully patched? Is the firewall configured properly? Is the system running unnecessary services? Information about these configuration details are scattered throughout the system, and are thus difficult to locate; even when users do track them down, they are often unsure about what action to take to correct any threat. The TALC project, which stands for Threat Awareness, Learning, and Control, is developing an extensible tool that can perform real-time auditing of a range of aspects of system configuration (comparing, for example, installed software versions against CERT advisories). Identified network security weaknesses are then visualized directly and non-intrusively on the desktop as on-screen graffiti -users can perceive when their systems are becoming more and more at risk as the level of detritus rises. This tool also allows threat mitigation to be directly actionable -by clicking on the background, users are presented with an interface that allows them to remove the threat, "cleaning up" the on-screen graffiti.

Tiger Teams Student Design CompetitionUsable Security

The Pixi Lab is the host of the Tiger Team Student Design Competition for Usable Security. Held 2005-2007, this competition allows two-person student teams (representing both HCI and Security) to compete for funding to develop novel projects that address challenges of usable security. The competition has been supported generously by Symantec, IBM Internet Security Systems, and Google.

inSpaceTangible Interaction

The inSpace project is an interdisciplinary collaboration with Steelcase, and brings together researchers with experience ranging from human-computer interaction, interaction design, software development, industrial design, furniture and interior design, and architecture. A central goal in this research partnership is to understand how guiding principles in the design of the physical world should inform the design of the digital, and more importantly, how these two layers should be co-designed.

Our current space includes a number of technical artifacts that work together to support collaboration, including a table with integrated sensing and ambient feedback, a shared wall display supporting multiple simultaneous users, and a collection of storage and infrastructure services for communication, wused in the space.

WindowScapeInnovative Interaction

WindowScape is a task oriented window manager designed to help people access windows as groups, as well as organize and manage individual windows more easily. WindowScape allows windows to be miniaturized down to thumbnail size and organized by the user. This temporally stable layout of miniatures allows people to leverage their spatial memory when searching for a window. To enable people to switch between groups of windows, WindowScape builds a timeline of window configuration states, and provides techniques to rapidly return to those prior configurations. In effect, these history states act as implicit, post-hoc window groups. It is implicit because the user never has to explicitly choose whether a given window belongs in a particular set. These decisions are only indirectly made through the history states the user decides to access. The groupings are post-hoc because the user does not have to make a priori decisions about where a widnow belongs. They can instead defer the decision until the last moment when they actually need to use access the group. WindowScape also offers interesting techniques for accessing obscured windows while largely maintaining a stable display image, as well as enabling keyboard navigation among arbitrarily arranged objects.

http://www.cc.gatech.edu/~craig/windowscape.htm

uMiddle Interoperability FrameworkInnovative Interaction

uMiddle is a "bridging framework" for supporting interoperability among diverse platforms. Through its extensible architecture, uMiddle can allow dynamic interoperability among Bluetooth devices, Universal Plug and Play (UPnP) services, and a range of other platforms. We have used uMiddle successfully in a range of interactive applications, including tools designed to seamlessly bridge physical and digital representations of data.

End-user Perceptions of UbicompSocial Impacts of Technology

How can designers of ubiquitous computing technologies ensure that they understand the non-functional needs, values, and expectations of end-users? We used a qualitative method from public policy to elicit reflective feedback from end-users about technologies that they may not yet have used nor fully comprehend. Our study uncovers information about end-user perceptions of RFID, including a range of "folk theories" held by the public about this technology, and their associations of it with certain social groups and values. We argue that these perceptions can limit technological adoption, and conclude with a discussion of challenges for the design and deployment of ubiquitous computing systems.

Technology and HomelessnessSocial Impacts of Technology

Technology is all around us, and while we consider new and novel uses for mobile and ubiquitous computing it is also important to consider how these kinds of technologies affect members of our society who do not have access to them. We have begun looking at this problem by conducting an qualitative study of the homeless in Atlanta, specifically talking with them about their use and perceptions of technology. This study is the first piece of a larger body of planned work to explore how marginalized and underserved communities deal with a world increasingly dependent on technology.

Studies in Photo SharingSocial Impacts of Technology

Our group has empirically studied photo sharing practices, including the practices of users of the Flickr web site. Our studies reveal differences in the privacy orientation between traditional consumer snapshot photographers and Flickr users. Further, we explore how both online and offline practices are intertwined, particularly around Flickr users' "photo strolls"--coordinated real-world events centered around group photo taking, rather than the photo sharing events that have been observed previously. Our studies point to a number of suggestions for the design of future consumer photography tools.