Where? - Collaborative Web Search: Who, What, Where, When, and Why - page 33

co-located collaborative searches both in office and home settings; they found that while some co-
located searches were quite involved, lasting many minutes or hours, a majority were rather brief,
serendipitous events, lasting only a few minutes. Based on the findings of their study, they rec-
ommend that designers of systems for co-located searching keep system start-up costs minimal,
provide a history of a group's suggestions, enable distribution of control among group members,
include consensus facilities, provide awareness mechanisms and shared context, and make it easy for
each participant to take relevant information away from the shared setting.
Researchers have begun to explore how to support co-located search scenarios using the
different resources that may be available in different co-located settings (e.g., personal computers,
mobile devices, large displays). In this section, we first give an overview of the range of solutions
that have been explored for supporting co-located collaborative search. We then explore two such
systems, CoSearch and WeSearch, in-depth, in order to illustrate the unique design considerations
that must be considered when designing for co-located groups.
Systems for co-located collaborative search tend to include a separate input device for each
group member, oftentimes augmented by a shared display. For example, Cerchiamo ( Pickens et al. ,
2008 ) (Figure 2.2) supports co-located collaboration among a pair of users searching over a video
archive. Each partner uses his own PC, which shows the individual information relevant to that user's
distinct role in the search task (either entering query terms as the “prospector” or triaging results as
the “miner”). Cerchiamo supplements these personal computers with a large, shared wall display that
shows data relevant to both users (i.e., the overall progress of the search task). WebGlance ( Paek et al. ,
2004 ) (Figure 4.8) supports co-located Web browsing, although it doesn't focus on search per se .
Users can control a Web browser on a large, shared display, by sending commands from their PDAs.
Maekawa et al. ( 2006 ) (Figure 4.9) support visual search of a webpage by dividing a large page into
non-overlapping sections and sending one section to each group member's mobile device, enabling
a divide-and-conquer approach. Query By Argument ( Blackwell et al. , 2004 ) (Figure 4.10) is an
educational system focused on rhetorical argument construction, that provides students with RFID-
tagged tangible objects representing key concepts from a document collection; manipulating these
tangibles re-ranks the relevance of items in the collection, helping students to see which information
might support their current arguments. CoSearch ( Amershi and Morris , 2008 ) combines group
members' mobile phones with a shared display to support collaborative Web search; we discuss
CoSearch in-depth below.
While some of the aforementioned systems used shared displays, they all provided individual
devices (phones, PDAs, PCs, or tangibles) for each group member. However, as touch-sensitive
devices have increased in prevalence, researchers have also begun to explore how a single, shared
multi-touch display can support collaborative search (generally in a tabletop form-factor, so as to
enable face-to-face work). Morris et al. ( 2010a ) described the design space of surface computing
systems that support co-located collaborative search (Table 4.1). Their design space considers two
primary dimensions: the group's configuration (such as the size of group supported, the spatial
arrangement of group members, the ecology of devices they bring with them to the shared display,