Database Reference
In-Depth Information
functions, and uses reasoning techniques to infer service annotations.
44
Over
800 services are described using ontologies
45
expertly annotated by a full-time
curator used by clients such as Find-O-Matic, its discovery tool, Feta,
46
which
is only available as a plug-in from the Taverna Workflow Workbench. The Bio-
Catalogue
*
project
47
incorporates the experiences of the Taverna Registry and
myExperiment (Section 13.6) to build and manage a richly described catalog
of Web services in the Life Sciences. The catalog's services have descriptive
content capturing functional capabilities curated by experts and by the com-
munity through social collaboration; operational content such as quality of
service and popularity is automatically curated by monitoring and use anal-
ysis. The BioCatalogue is a free-standing component with its own RESTful
APIs that can be embedded within and accessed from third-party applica-
tions. Developers can incorporate new services through simple actions and can
load a pre-existing workflow as a service definition within the service palette,
which can then be used as a service instance within the current workflow.
Taverna also supports the configuration of the appearance of the graphical
representation of workflows, so that a workflow can be suppressed to give
higher-level views, for example, to remove details such as data translation (or
other “shim”) services.
†
One of the most powerful aspects of Triana is its GUI. It has evolved in
its Java form for over 10 years and contains a number of powerful editing
capabilities, wizards for on-the-fly creation of tools, and GUI builders for cre-
ating user interfaces. Triana editing capabilities include multilevel grouping
for simplifying workflows, cut/copy/paste/undo, ability to edit input/output
nodes (to make copies of data and add parameter dependencies, remote con-
trols, or plug-ins), zoom functions, various cabling types, optional inputs, type
checking, and so on. Since Triana came from the gravitational-wave field, the
system contains a wide-ranging palette of tools (around 400) for the analy-
sis and manipulation of one-dimensional data, which are mostly written in
Java (with some in C). Recently, other extensive toolkits have been added for
audio analysis, image processing, text editing, for creating retinopathy work-
flows (i.e., for diabetic retinopathy studies), and even data mining based on
configurable Web services to aid in the composition process. See Taylor
48
for
a further discussion and description of such applications.
Wings
35
uses rich semantic descriptions of components and workflow tem-
plates expressed in terms of domain ontologies and constraints. Wings has
a workflow template editor to compose components and their dataflow. The
editor assists the user by enforcing the constraints specified for the work-
flow components. It also assists the user with data selection, to ensure the
datasets selected conform to the requirements of the workflow template. With
*
http://biocatalogue.org
†
Shims align or mediate data that is syntactically or semantically closely related but not directly
compatible.
49