Information Technology Reference
In-Depth Information
metaphor is one of the most explored (e.g. [14][15][16][17]). For example,
in [18] Wettel and Lanza propose a city metaphor for the comprehension
of object oriented software systems. Similar to our proposal, the metaphor
proposes a large and fine-scale understanding of software. In addition,
familiar things are used for the visualization of software systems: classes
are represented as buildings and packages as districts. The metaphor is
implemented in the CodeCity tool; and to improve the realistic aspect of
the city, authors focused on the design of the urban domain. Our metaphor
is different because the fine-grained representation of classes takes into
account a larger number of metrics (e.g., number of public methods and
lines of source code commented). However, we plan a users study to
understand whether the use of more information results in a deeper
understanding of the system. In a recent paper [19], the same authors
present a controlled experiment with professionals to assess the validity of
their metaphor and supporting tool. The results show that the CodeCity
tool leads to a statistically significant improvement in terms of task
correctness. The results also indicate that the use of the tool statistically
decreases task completion time.
Graham et al. [20] propose a solar system metaphor where suns
represent packages, planets are classes, and orbits represent the inheritance
level within the package. Such a metaphor is used to analyze either static
or evolving code in order to perceive in real time suspected areas of risk
within the code base. Again, the most remarkable similarity is the use of a
metaphor familiar to the maintainers. The most significant difference is
that our fine-grained representation allows the maintainer to get more
information on each class of the subject system.
Kleiberg et al. [21] propose a botanical tree metaphor which is
different to ours. The authors suggest forests of trees for the visualization
of huge hierarchical structures and apply the proposed metaphor to the
visualization of directory structures. Directories, files, and their relations
are visualized using trees. The approach is basically a natural visual
metaphor for hierarchically structured information. In contrast, we use
trees as fine-grained representations of classes to create a large-grained
representation, in which each sub-forest contains classes of the same
package.
Martínez et al. [22] suggest a metaphor based on landscape, whose
main objective is to visualize the integrated representation of software
development processes. The metaphor is conceived to describe several
aspects of the development process. Their metaphor does not provide a
means to visualize source code.
Search WWH ::




Custom Search