When? - Collaborative Web Search: Who, What, Where, When, and Why - page 50

These practices are similar to what researchers have observed people do to keep information
they have encountered around for future personal use ( Jones et al. , 2002 ). In a way, the keeping
of found information is like collaborating with one's future self, with the individual sharing links
and information that will help that individual pick the task up later ( Morris and Horvitz , 2007a ).
Martha, for example, could have emailed herself the links she found while searching immediately
after receiving her diagnosis that she thought would be useful to her later. As noted by Jones et al.
( 2002 ), a number of the methods people use to make things easier to find later actually have social
or communication benefits. Although Jones et al. studied diverse keeping methods, such as book-
marking or saving to a file, from a personal re-finding perspective, the authors highlighted emailing
information to other people, printing out information, and posting information to a website, all
mentioned above, as having additional communication benefits.
The tools designed to help people manage their complex individual search tasks that extend
over time include bookmarking, histories, or systems that allow users to flag pages or parts of pages for
inclusion in a workspace, such as SearchPad ( Bharat, K. , 2000 ) or Hunter Gatherer ( Schraefel et al. ,
2002 ). These systems allow an individual to collaborate with their past self when they resume a
previously begun search task.
Social search systems that allow individuals to take advantage of information from others,
but that are non-collaborative, are generally asynchronous in nature. For example, data from a
community of related users running searches at earlier times can be used to re-rank search results
(e.g., ( Smyth, B. , 2007 ; Smythetal. , 2009 ; Teevan et al. , 2009b )). And systems in which a user's
question is federated to a social network (e.g., ( Evans et al. , 2010 ; Morris et al. , 2010c )) tend to
receive responses after a period of time has elapsed. Understanding how to manage the trade offs
between the speed of responses from traditional search engines and the personalized, high-quality
responses returned asynchronously from social sources is an open research problem.
For systems supporting asynchronous active collaboration among users with shared goals, the
creation of a persistent artifact representing the current state of the search is important. This artifact
may exist as stateful data stored in the cloud (e.g., as in SearchTogether ( Morris and Horvitz , 2007b ))
so that it can be accessed by group members at any time. Alternately, a file metaphor can be used,
enabling sharing of the artifact via traditional mechanisms such as email. S 3 ( Morris and Horvitz ,
2007a ) is an example of a collaborative search system specifically designed to support this style of
asynchronous collaboration; we discuss S 3 in more detail in the following section.
5.1.1 EXAMPLE: S 3 (STORABLE, SHAREABLE SEARCH)
The S 3 (“Storable, Shareable Search”) system facilitates asynchronous collaborative Web search tasks
by adopting a file metaphor. S 3 uses a persistent XML data format (referred to as a .srch file) to
represent both the process and products of an ongoing search task. This file can be saved, loaded,
edited, and shared in a manner analogous to a word processing document or spreadsheet. The
opening, saving, and editing takes place using the special S 3 browser (Figure 5.1). Sharing occurs by
sending the resulting .srch file as a simple email attachment.