Documentation

What are Workflow-Driven Ontologies (WDOs)?

• WDOs are Task Ontologies: According to Guarino (1997), ontologies can be categorized according to their level of dependence on a particular task or point of view. For example, Top-level Ontologies describe general concepts, e.g., space, time, and matter. Domain Ontologies specialize terms used in top-level ontologies to describe vocabulary of a particular domain, e.g., crust and moho in the domain of geophysics. Task Ontologies use vocabulary described in a domain ontology to describe tasks or activities of that domain, e.g., creating a crustal model of Earth.
• WDOs are OWL-encoded ontologies: The OWL Web Ontology Language is an XML-based language recommended by the W3C to encode ontologies using Web conventions. In other words, the concepts and relations defined in an OWL ontology are encoded in a language that can be transmitted over HTTP and have a URI that can be used to make them web-accessible.

Consider the following example about the classification of ontologies and the relation between them:

1. The WDO-upper level ontology is a Domain Ontology that defines concepts such as Data and Method. This ontology is published on the web and can be found at http://trust.utep.edu/1.0/wdo.owl. As a requirement, all WDOs need to reference (or import) this ontology in order to be considered a WDO.
2. The Geophysics ontology is a Domain Ontology that defines concepts from the domain of Geophysics. For example, a concept such as Crust could be defined here.
3. The Crustal Modeling WDO is a Task Ontology that defines concepts regarding the activities about creating crustal models of the Earth.
4. The arrows denote dependencies between ontologies. For example, in order for the Crustal Modeling ontology to be denominated a WDO, it has to reference and use the concepts in the upper-level WDO. Optionally, the WDO may reference other ontologies that define vocabularies that are more established in the field, such as an ontology about the domain of Geophysics. The action of reusing concepts from other ontologies is called harvesting.

An example of a WDO encoded in OWL can be found at http://trust.utep.edu/2009/gridding/wdo and the WDO-It! tool provides a GUI to view and edit it.

What are Semantic Abstract Workflows (SAWs)?

• Data flow models of a process: SAWs are graphical structures that model data flow of a process. The graphical structure contains three types of entities: 1) Rectangle-shaped nodes represent activities of the process, 2) oval-shaped nodes represent data sources and sinks, and 3) directed edges connecting the nodes represent the data of the process. The connection and direction of an edge determines whether data is coming from a source, is going to a sink, is input to an activity, or is output from an activity. Furthermore, edges link nodes to represent data dependency between them.
• OWL-encoded knowledge bases: Just like ontologies, OWL can be used to encode knowledge bases. SAWs are considered knowledge bases because they do not contain concept definitions. Instead, SAWs contain instances of concepts defined in ontologies (i.e., WDOs).

An example of a SAW encoded in OWL can be found at http://trust.utep.edu/2009/gridding/gridDataset and the WDO-It! tool provides a GUI to view and edit it.

Why do I need WDOs and SAWs?

• To document existing scientific processes
• To design new scientific processes
• To establish common vocabularies about a scientific process
• To understand scientific processes of others

How do I create my WDOs and SAWs?

• WDO-It!

Using WDO-It!

Terminology

• Class: Same as Concept.
• Concept: Concepts are general entities that embody common features of a particular set of instances. Concepts are intended to facilitate the representation of knowledge by providing a grouping mechanism. For example, the concept Car may be used to describe all the entities that have four wheels and two or more doors. If the concept Car was not used, the altenative to represent knowledge about which entities are cars would be to list them all.
• Data:
• Harvest:Term used to describe the activity of reusing ontological concepts. By 'harvesting' concepts from more established ontologies in a domain, the intention is to avoid reinventing terms and to achieve consistency in the terminology being used. Note that the ontology creator(s) is (are) ultimately responsible to determine the appropriateness of harvested concepts in the context of the ontology being created.
• Import:This is a technical term related to OWL.. By including an tag in an OWL document with a reference to the URI of another OWL document, the intention is to access the concepts, relations, and instances defined in the referenced OWL document. This is useful, for example, to harvest concepts from other ontologies, or to separate ontologies from knowledge bases (i.e., by defining ontological concepts in one OWL document and referencing the concepts from another OWL document to define instances).
• Instance: These are entities that represent individuals of a class or concept. In fact, instances are called individuals in OWL. For example, a concept (or class) may be Car, and an instance of Car may be John's Car; furthermore, a second instance of the concept Car may be Mary's Car. Instances are said to be of a specific type, where the type is the class or concept that they belong to. To follow the previous example, both instances of John's Car and Mary's Car have the type Car.
• Knowledge base:
• Method:
• Ontology:
• OWL:
• RDF:
• Sink:
• Source:
• URI:Uniform Resource Identifier
• Workflow:
• Workspace:

Overview of the GUI

1. Loaded OWL Documents tree
2. Concept hierarchy for selected ontology
3. WDO Data concept hierarchy
4. WDO Method concept hierarchy
5. SAW scratch area

HOW TO

WDO-related tasks

Create a WDO

Create a WDO concept

Remove a WDO concept

Add a comment to a WDO concept

Rename a WDO concept

SAW-related tasks

Create a SAW

Create an instance in a SAW

Remove an instance in a SAW

Assign a name to an instance in a SAW

Connect a Data instance to a Method instance

Disconnect a Data instance from a Method instance

Make two Data instances originate from the same Source

Make two Data instances end in the same Sink

Specify more process detail about a Method instance with a second SAW

Other tasks

Create reports of WDOs and SAWs

Create data annotators