![](bg3.png)
concept is prevalent throughout the design of the in-
terface.
2.3 Mind Mapping of Workflow
Mind Mapping has been used in various fields to
help categorize ideas and give a visual flow be-
tween concepts (Brinkmann, 2003). The process
of mind mapping allows both sides of the brain to
work together to increases productivity and creativ-
ity. Mind maps are great tools for organizing infor-
mation. “The hierarchical structure of a mind map
conforms to the general assumption that the cognitive
representation of knowledge is hierarchically struc-
tured” (Brinkmann, 2003) . This lends itself well to
the notion of a compositional system. The user will
lay out and connect the compositional units necessary
to visualize their solution.
Mind maps excel in connecting new information
with given knowledge (Brinkmann, 2003). This qual-
ity is incredibly useful when trying to expand a work-
flow in an ODCS. Often new algorithms or computa-
tional units will become available, and the user may
choose to add them. By having a visual workflow it
will be easy for the user to identify the location of
where the new CU should be placed.
The mind mapping process is a logical way to rep-
resent a workflow as it gives coherence and a good
visual representation. It is evident with some of the
design choices in the interface presented that mind
mapping was a key inspiration allowing the interface
to truly leverage the knowledge of the user.
2.4 Related Existing Systems
With the acknowledgment of the different user types
and their skill levels, many research initiatives have
been started with the goal of meeting some of the
heuristics discussed in section 2.2. In this section we
provide an overview of some of these initiatives.
Triana is a problem solving environment (PSE)
that can be used for composing, compiling and run-
ning applications (Majithia et al., 2004). Like many
other compositional systems, the goal of Triana is to
make use of several tools (data analysis tools, algo-
rithms, control structures) to solve a problem (Tay-
lor et al., 2007). Triana defines interfaces to a vari-
ety of execution environments through its Grid Ap-
plication ToolKit (GAT) and Grid Application Proto-
col (GAP) allowing the execution of both task-based
workflows and service-based workflows. With Tri-
ana, a user graphically composes a workflow by dis-
covering, composing, invoking and publishing com-
posite services in a seamless manner.
Galaxy is a tool and data integration framework
(Team, 2010). It is an open-source application and al-
lows for the installation of individual instances. The
target users in Galaxy are software developers and bi-
ologists. The tools found within the application can
be applied to datasets to perform calculations. New
tools can be added to the interface. In order to do so,
a developer must add a configuration file that contains
information about how the tool is run (Team, 2010).
One form of current use is as a web-based genome
analysis tool (Goecks et al., 2010). The genome anal-
ysis is done in the workspace area which allows for
the application of computational tools to perform the
data analysis (Goecks et al., 2010). It is important to
note that this application targets a specific field, bi-
ology. Furthermore, in order to keep the front-end
user satisfied, it is required to maintain a person with
some programming experience in case new tools are
needed. Another aspect of Galaxy worth noting is
that although it is a web-based application, installa-
tion only works through Linux and Mac OSX, with
no current support for Windows (Team, 2010). These
points play a role in increasing the cost and decreas-
ing the adaptability of the application by other users
or fields of interest.
Yahoo Pipes is a composition tool that creates data
mashups; it is used primarily as a simple way to run
web projects, or publish web services (Pruett, 2009).
Yahoo Pipes has a set of modules that have different
types of containers that can be used when building
the pipe. It contains many of the concepts that are
useful for CU workflow composition, however the set
modules that it contains would have to be configured.
The power of ontologies is that they have the abil-
ity to automatically infer semantic relationships be-
tween CUs within the ODCS. For example, by de-
scribing the input/output parameters for each CU, the
system would have the ability to infer which other
CUs have the same I/O requirements. It would also
have the capability of determining which CU’s were
created by the same software developer. In turn, this
reduces the amount of work on the software develop-
ment side thus reducing any dependence the front-end
user might have when making modifications, such as
adding new CU’s, to the system.
The discussed systems are efficient in their areas
and have a lot of power, but all require users with
experience and background knowledge in program-
ming in order to make modifications to fit their needs.
When considering the perspective of the end user,
some of the described systems are not very adaptable
to change. By using an ontology-driven system, the
user is able to intuitively run the system, and perform
their necessary actions without being required to at-
KEOD 2011 - International Conference on Knowledge Engineering and Ontology Development
178