Information Technology Reference
In-Depth Information
process of generating a browser from this model. To demonstrate the approach,
we describe how a web application to browse data about researchers and their
publications has been implemented.
We begin in Sect. 2 with a more detailed discussion of the background to this
work and related research before going on to describe our approach in Sect. 3.
Sect. 4 then introduces the SQL extension used to define the view model and
shows how it can be used to specify a browser for a particular web search service.
Sect. 5 provides details of the system architecture and the process of generating
a browser from the view model definition. Implementation details are then given
in Sect. 6. We discuss the contributions of our work in Sect. 7, and concluding
remarks are given in Sect. 8.
2 Background
Tag clouds have become an extremely popular way of providing visual summaries
of data collections and are nowadays used in many Web 2.0 sites to provide
a basic search service based on user-generated tags. For example, both Flickr
1
and Del.icio.us
2
provide search services based on collaborative tagging. Although
very simple, tag clouds can be used to support search, browsing and recognition
as well as forming and presenting impressions [1,9]. In previous work, we have
shown that tag clouds can also be used as the basis for a generic database
browser [6].
The presentation and layout of tags can be controlled so that features such as
the font size, type and colour can be used to give some measure of the importance
of a given tag, while the positioning of tags may be based on pure aesthetics,
alphabetical sorting or some form of relationship between tags. Studies have
experimented with such features and their impact on users, concluding that font
size, font weight and intensity are the most important features [7,1]. A study on
search performance [8] found that topic-based layouts produced better results
than random arrangements, but alphabetic layouts were best.
More recently tag clouds have been proposed as a means of summarising and
refining the results of keyword searches over structured as well as unstructured
data [5,3,4]. In [5], tag clouds are used to summarise query results of the PubMed
biomedical literature database based on words extracted from the abstracts re-
turned by a query. The term
data cloud
is used in [3,4] to refer to their particular
adaptation of tag clouds for summarising keyword search results. Data clouds
were implemented as part of CourseRank, an application that enables students to
search for classes, give comments and ratings, and also organise their classes into
a personalised schedule. The developer of a data cloud application specifies how
application entities can be composed from the relations in the database in or-
der that keyword search can be applied to entities rather than simple attributes
or tuples. The keyword search is based on a traditional information retrieval
1
http://www.flickr.com
2
http://www.delicious.com
