News

TopQuadrant Releases Semantic Web Dev Tools

• By Kurt Mackie
• January 9, 2008

This collection of tools is part of the company's TopBraid Suite of solutions that support ontology modeling, Web application deployment and collaboration. TopBraid Composer Maestro Edition has enhanced features for testing and rapid deployment of Web applications, according to the company. It includes an integrated Web server and Business Intelligence and Reporting Tools (BIRT), plus ease-of-use improvements.

Maestro Edition users can execute Java Server Pages (JSP) in the Eclipse development environment to generate HTML and XML documents. The toolset has a mapping feature that lets users query and transform XML in Resource Description Framework (RDF), and then save back to XML without losing information (which is known as "round-tripping"), according to company officials. Users can also scan e-mails and load metadata into Web Ontology Language (OWL) ontologies.

The key idea behind the Maestro Edition is its ability to enable agile development, according to Holger Knublauch, vice president of product development.

"We offer an integrated solution that provides coverage for the whole development lifecycle, starting from design to deployment," Knublauch said. "The new feature here in Composer Maestro Edition is the ability to use the development environment as a testing server, as a testing framework. So essentially the development lifecycle is very agile, with short development and deployment cycles."

Knublauch is an expert in semantic Web technology based on his work with Protégé-OWL, which is open source software that first became known in academic circles. He brought his expertise in semantic Web technology to TopQuadrant, a privately held and employee-owned company largely made up former IBM Global Services employees. TopQuadrant, formed in 2001, has offices in Alexandria, Va. and Mountain View, Calif., plus a subsidiary in Seoul, South Korea.

TopQuadrant's products use a proprietary scripting language called Sparql/Motion, which is built around the W3C's SPARQL query language. SPARQL is comparable with SQL, but it operates on RDF data sources, according to Dean Allemang, TopQuadrant's chief scientist.

Knublauch described Sparql/Motion as "a visual scripting language for the semantic Web, comparable to Yahoo Pipes." The query engine lets people define small scripts in a graphical editor, he explained.

"They can take information from a new seed, database, and some other source, integrate them and then run some transformations on the data," Knublauch said. "You can display things on a map or a calendar or just create another database or another file. We are using OWL itself to express those scripts."

So what is the semantic Web and why would companies need it? Allemang defined it as "a Web of data." The semantic Web uses ontologies that can be understood by humans but also can be read by machines for things like search functions to better access data on the Web or better facilitate interoperability. The semantic Web helps find things through URIs (Uniform Resource Identifiers), and Allemang explained how that works using an analogy from the science of linguistics, which has a similar concept called "referential semantics."

"[With] referential semantics, a word or utterance has to stand for something, and knowing what your symbol stands for is the referential semantics of that symbol," Allemang said. "And that's the sense in which the semantic Web is semantic. You have these symbols, which are actually called URIs, and these URIs stand for something. And the semantic Web is about not just having documents, not just having data, but actually having a symbol and the thing for which it stands."

Business is interested in semantic Web technology because of its multiple data sources and the problem that legacy data can't be easily combined with new data sources, Allemang explained.

"One way to look at this is that [companies] have been putting all of this information in Web applications for the last 10 years and at some point they are going to want to get it out," he said. "It's very difficult to do that in a sensible way. It takes a lot of programming effort to take a legacy application and adapt it to do business processes."

Can semantic Web technology avoid the problems of trying to store relational data and hierarchical XML data, and not lose information in the process?

"What semantic Web triple-store vendors are doing is getting their triple stores to scale up to the same kind of scales that XML stores and relational stores do," Allemang explained. "And mathematically, one of the nice things about RDF is that it is the simplest building block that you can build data from. So it means you can represent XML and relational tables in RDF -- both of them and in the same store. So it's actually now possible, if you go to RDF, to take XML and dump it into RDF, take relational data and dump it into RDF, and start talking about how this XML node relates to that cell in that relational database and have it make sense. Now, there's all sort of technology around what I just said, but that is the sort of vision that is driving a lot of the data vendors in the semantic Web world right now."

TopQuadrant doesn't produce a database to make this happen. Rather, the company offers a framework to bring in RDF data stores. Allemang said that one the premier RDF data stores is offered by database provider Oracle.

"So, Oracle is getting into this," he said. "As far as we know, IBM and Microsoft haven't entered this space in a big way. Oracle did a product announcement about a year and a half ago. For Oracle 11g, we can now access it through TopBraid. One of the key things you can do with TopBraid Suite is you can choose which of these things you want -- do I want Oracle, do I want AllegroGraph, do I want some of the open source ones? -- you get to pick which one."