Now, changes to applications can be made without restarting the server, said Holger Knublauch, vice president of product development at TopQuadrant. This, in turn, reduces QA cycles, he said.
Semantic development is an emerging field that may be viewed as a subset of data integration. Interest in semantic development was largely spurred by Tim Berners-Lee, inventor of the World Wide Web and now head of the W3C, where he has promoted the idea of the "Semantic Web." Berners-Lee and W3C have developed a Resource Description Framework (RDF), related schema, and the Web Ontology Language (OWL) to enable software agents to work within specific knowledge domains.
TopQuadrant offers a TopBraid Composer development suite for modeling semantic applications. Knublauch, the original developer of the Protégé Web Ontology Language plugin, said one difference between semantic Web application and mainstream conventional software is that much of the behavior is model-driven.
"For example, the UI may display certain forms and constrain the input by certain values, depending on the structure of the underlying ontology," he said. [An ontology as described here may define a set of representational primitives for modeling a domain of knowledge.]
Knublauch also said the application logic may be influenced by rules and logic defined in the ontologies.
"These ontologies are often open-ended networks of other ontologies and new information from all over the world, mashed together dynamically," he said. "Semantic Web application execution engines, therefore, need to be well-designed to be able to handle 'unforeseen' situations."
Knublauch said TopBraid components are prepared to deal with diverse scenarios. He added that the TopBraid Composer provides various generic testing capabilities such as SPARQL libraries, rules, classification and consistency checking.
"However, a good deal of work is also [in] experimenting," he said, "and this is where rapid turn-around times are crucial."
To a large extent, the type of data that semantic development will focus on is unstructured text. The sources of data are varied, and they may reside on different systems and templates, and use diverse naming rules. An example of a semantic application would be a system that could spot the early signs of an influenza outbreak by reviewing health and epidemiological data from caregivers, hospitals, pharmacies and other sources. Such a system, in fact, was created in the wake of Hurricane Katrina.