Poster abstract for the Tabletop 2007 workshop, Newport, Rhode Island, USA, October 2007
Designing Tabletop Interfaces for Asymmetric Distributed Collaboration
Mark Ashdown1, Stacey D. Scott1,2
Humans and Automation Lab, Massachusetts Institute of Technology, Cambridge, MA,USA
Systems Design Engineering, University of Waterloo, Waterloo, ON, Canada
Abstract They must remain aware of the state of all relevant
personnel and resources, and consider information
Complex task domains such as emergency response from many sources.
and command and control often involve collaboration These different task activities and environmental
between operational personnel in the field and tactical constraints lead to different technological requirements
personnel in a central command centre responsible for for any support systems developed for these tasks.
coordinating the efforts of those operational Operational personnel, who tend to be mobile and
personnel. The asymmetry in their respective work can experience physically demanding situations, would
environments, job responsibilities, available need an extremely portable and robust device, such as a
information, and situation constraints produce small handheld computer. In contrast, tactical
distinctly different technological requirements for personnel have few such environmental constraints,
potential support systems for these different personnel. and thus can exploit stationary, large display systems.
This research focuses on the use of a tabletop display Large displays, such tabletop displays, provide benefits
to support the planning and coordination duties of the for productivity  and spatial awareness .
tactical personnel. A primary goal is to address the Additionally, a tabletop system could display the maps,
inherent challenges of designing large-screen tabletop schedules, documents, and other information necessary
interfaces that support synchronous interaction, data to support the tactical role. Advances in networking
sharing, and coordination with remote collaborators will enable data sharing between the devices used by
who have significantly diminished technological the operational and tactical personnel, in addition to
capabilities, particularly in terms of available display standard voice communication (Figure 1).
1. Asymmetric Distributed Collaboration display
Various domains, including emergency response
and command and control, involve coordinating a command centre
distributed team of people from a command centre. field
Operational personnel are those people situated in the operational
field, responsible for performing the physical work handheld
such as gathering information or material items, or displays
providing services. Tactical personnel are those people Figure 1. Asymmetric distributed collaboration.
in a command centre coordinating the actions of We are designing interfaces for linked tabletop and
multiple operational personnel. The difference between handheld devices, to support real-time collaboration in
these roles leads to several forms of asymmetry. a shared visual space. We will study how people
Operational personnel execute physical tasks (often collaborate in a test scenario to characterize the
while traversing the field setting), have detailed techniques they use and to inform design
knowledge of the immediate situation around them, and recommendations. In particular, we are interested in
receive information about new tasks from the tactical how the collaborators deal with the asymmetry in
level. In contrast, tactical personnel perform the higher- hardware and roles, which communication strategies
level functions of planning for future tasks, and they use to deal with the mismatch of information
coordinating the various operational personnel to visibility, and how activity awareness between
ensure that the team satisfies the overall mission goals. collaborators should be supported.
2. Design Challenges scenario. Figure 2 shows an extract of this diagram for
the tactical coordinator’s scheduling work. The
Many existing design approaches for supporting diamonds indicate decisions and the boxes indicate
distributed collaboration assume fairly symmetric processes. Thick-edged boxes indicate that
situations between participants in their work collaboration with a search unit is required to complete
environments, technological capabilities, and their the process.
roles and responsibilities during collaboration. In monitor units
asymmetric collaboration, however, people tend to
have complementary job roles with different
technological requirements. Research indicates that search unit N
Y external report N
or structure status victim found?
such role asymmetry and platform heterogeneity can change?
negatively impact remote collaboration ; thus, obtain timing obtain victim
careful design is required for our envisioned system of change info.
schedule change N report
networked tabletop and handheld computers. Providing
Y modify rescue
a shared visual space between distributed collaborators modify search
facilitates communication by supporting conversational schedule
grounding. However, a vast difference in display size Schedule inform units of
provides an obvious challenge to offering standard new schedule
shared visual solutions, such as a WYSIWIS (What- Figure 2. Event flow for the tactical actor.
You-See-Is-What-I-See) interface design. Thus, new
visibility techniques may be required to support The CTA results in a set of information and
workspace awareness . functional requirements, but leaves considerable scope
To develop technologies that address these issues, for the specific design of the interface. Since the large
we are developing an experimental platform involving horizontal surface and direct input of a tabletop display
a representative asymmetric distributed collaboration make it qualitatively different from conventional
task scenario: urban search and rescue (USAR). interfaces, new interaction techniques may be required,
especially for supporting synchronous collaboration
3. Urban Search and Rescue Scenario with maps and schedules.
Our USAR scenario involves two types of 5. Conclusion
distributed collaborators: search units (operational
personnel) and a tactical coordinator. The search units Our poster will present initial CTA results, focusing
navigate around a city, searching buildings and on the design requirements for the tabletop display. We
reporting victims found and unexpected events. They hope to gain feedback from the community on the
will have handheld displays with which to view maps implications of these requirements for the design of the
and task schedules, and they may speak to the subsequent tabletop interface, and to discuss more
command centre via a voice channel. The tactical general issues regarding using a digital tabletop to
coordinator is situated in the command centre. This support asymmetric distributed collaboration.
person monitors and coordinates the search units, in
addition to other units that comprise the team. The
tactical coordinator will have a tabletop display to  M. Czerwinski, et al., “Toward Characterizing the
support the coordination of the team’s spatial activities Productivity Benefits of Very Large Displays”, Proc. IFIP
(with maps) and their temporal constraints and progress Interact 2003, pp. 9–16.
(with schedules and timelines).  C. Gutwin, S. Greenberg, “A Descriptive Framework of
Workspace Awareness for Real-time Groupware”. Computer
4. Designing a Tactical Tabletop Display Supported Cooperative Work, 11:3–4, 2002, pp. 411–445.
 C.E. Nehme, et al., “Generating Requirements for
To identify the display requirements for the Futuristic Heterogeneous Unmanned Systems”, Proc. HFES
operational and tactical support systems, a cognitive 2006.
 D.S. Tan, et al., “Physically Large Displays Improve Path
task analysis (CTA) specifically designed for futuristic
Integration in 3D Virtual Navigation Tasks”, Proc. CHI
task scenarios , was conducted. A key part of this 2004, pp. 439–446.
CTA process is the creation of an event flow diagram  M. Velez, et al., “Who's in Charge Here?
that represents the temporal constraints between the Communicating Across Unequal Computer Platforms”,
decisions and processes that may occur during the task ACM Trans. CHI, 11:4, pp. 407–444.