Biology Reference
In-Depth Information
level of abstraction and develop a system which automatically/semi- automatically
creates the underlying structure. Ontology-based approaches are aiming at such
an abstraction, allowing them to operate on the basis of custom concepts. Olingo is
an ontology-based approach we have developed to generate the underlying data
structures for PFTs. Olingo is developed in two flavors: a Web portal (Olingo Web
Tool (OWT)) and a plugin for the Protege ontology editor.
7
We use an ontology-based approach because ontologies are extensible mechanisms
typically used for describing, via human-readable text, the domain of disclosure,
which enables common understanding among scientists and software applications.
Although ontology provides an abstract view, it is suitable as input for the generation
of underlying structures. Thus, to avoid both schematic and semantic heterogeneity,
Olingo targets the generation of a number of explicit output formats for the applica-
tion cases from different scientific domains. It is the link between ontologies and
data structures. Application domain experts may use Olingo during the design phase
to generate the data structures needed for the creation of PFTs, which will subse-
quently be applied during the execution and the analysis phase of the experiment
by less experienced users.
It is evident that, in some simple and specific cases, a basic model can be used by
a proprietary generator to create the database schema that is later integrated within
the scientific application. Usually implemented in an ad-hoc fashion, through scripts or
parser generators, this approach does not scale well for large-scale scientific projects.
OWT is a Web application that, based on an ontology, contributes to generate
the underlying data structures for PFTs. Ontologies are extensible mechanisms
typically used for describing, via human-readable text, the domain of disclosure, which
enables common understanding among scientists and software applications. OWT
targets the generation of a number of explicit output formats for the application
cases from different scientific domains. It is the link between ontologies and data
structures.
Olingo produces five different output formats, including (1) relational schema
with a data definition language for relational databases, (2) Java classes providing
the source code of data structures, (3) XML Schema with a specification for XML
documents, (4) mapping files for two different frameworks (Castor and Hibernate)
that support persistence of Java objects.
7.5.3
Work fl ow Management System
The WS-VLAM workflow management system
8
is composed of a workflow editor
and a workflow enactor. The workflow editor completes the PFT by providing
support for the execution phase of an experiment. In the current implementation,
the execution model of the workflow is data-driven. Workflow components are
Search WWH ::
Custom Search