Coordinated Multiple Views for Exploratory GeoVisualization - Geographic Visualization: Concepts, Tools and Applications - page 39

3. each entity has temporal and chronology information, such that some coordinations
persist while others are temporary for a task;
4. each coordination is defined in a scope - for example it may be that the entities are coupled
for a specific task and then uncoupled when the task has finished;
5. the coordination has a certain granularity - for example, some entities may be coordinated
point-to-point while for others it may be possible to chain multiple entities together;
6. the coordination of entities may be initiated by the user or automatically by the system;
and finally,
7. there are different methods to update the information when entities are coordinated -
for instance, cold linking permits elements to be coupled once, warm linking permits
users to decide when the views are updated, and hot linking is when the information is
dynamically updated (Unwin, 2001).
There are three principle coordination architectures: (1) constraint-based programming
(Mcdonald et al. , 1990); (2) a data centric approach taken from the database community,
where change to one component of a relational database is tightly coupled to other compo-
nents (North and Schneiderman, 2000); and (3) the module view controller pattern, where
one component observes the model and updates the view when changes are made (Pattison
and Phillips, 2001; Boukhelifa et al. , 2003). Tools such as Waltz (Roberts and Waltz, 1998)
and Improvise (Weaver, 2004) implement many of the ideas of the rudiments. However, few
CMV developers incorporate comprehensive coordination capabilities in their systems.
3.4.7 Challenges for interaction and manipulation
Most exploratory systems allow the visualization to be adapted and viewed again from a
new parameterization, whether this information replaces the old or is displayed in a new
window. However, there are still many challenges. First, not many systems allow the user
to overlay the information. One overlay strategy is to generate the difference between the
two parameterizations (Suvanaphen and Roberts, 2004). This can be extremely useful as
the visualization would clearly show what has changed. However, few systems supply this
functionality. Second, it is not clear as to how many views are actually useful. For instance,
a user may easily get lost in a explosion of overlapping windows, hence there is certainly
a trade-off to using replacement, replication and overlay. Consequently, more research is
required to evaluate how many views are useful and to provide guidelines to when each
strategy should be used. Third, with large explorations the user can easily forget which view
represents what parameterization. Thus methods to support the user in their exploration
are required.
Both interactive filtering and interactively adapting mapping parameters are important
exploratory techniques and it is clear that much research has been achieved. However, many
of these techniques are not implemented in modern exploratory tools. Developers seem to