Jennifer Zaino
SemanticWeb.com Contributor

When Prof. Jerry Boetje began a student project at the College of Charleston's Computer Science department to develop CLforJava, the effort was designed to throw about-to-be graduates into the deep end of complicated problems with the kinds of tools and processes they would be dealing with when they got into the industry.

Six years and 150 students later, the software language platform for developing solutions for hyper-complex systems could prove to have applicability for the development of the Semantic Web.

A core part of the college's Software Engineering Capstone Project course, the work has focused on creating a new version of the Common Lisp language that runs on the Java Virtual Machine and is intertwined with the Java language in such a way that users of Lisp can directly access Java libraries and vice versa. In the late stages of development now -- close to full-scale public open source status -- the product was exhibited at the International Lisp Conference at the Massachusetts Institute of Technology (MIT) at the end of last month. It's only in the last year or year and a half, however, as the system has gotten closer to its final version, that its potential usefulness to the semantic web has taken shape.

"We are seeing other versions of Lisp that are getting in and working very well in these complex kinds of systems. As we get to their level we also bring this ability to integrate very easily and trivially with Java," says Boetje. "That's what we bring to the table."

Several companies, Boetje says, write Semantic Web software in Common Lisp. Lisp programmers can build ontologies in RDFS (RDF Schema) and OWL (Web Ontology Language) using semantic technology solutions. AllegroGraph RDFStore from Franz Inc., a leading provider of Common Lisp products, for example, is a modern, high-performance, persistent RDF graph database that can scale to billions of triples.
Unlike the Semantic Web, Lisp has been around for decades -- it's more than fifty years old. "One of the things about Lisp is it is extremely good at handling complex systems, particularly where you verge on AI (artificial intelligence) and things like that," says Boetje. Lisp was the language behind the first wave of AI applications, and today it powers complex systems such as online travel site Orbitz and some maintenance schedules for the International Space Station.

"But as we get into the reality of how the semantic web is built in the core of things, a lot of different languages will [participate in] it, including Java." That calls for the creation of a Lisp system that can transparently get to Java if it needs to, and otherwise do work on handling the very complex issues of the semantic web, he says.

Take, for example, dealing with enabling enterprise application interoperability along the lines of services-oriented architecture. "We do know a lot of this architecture can benefit from AI (artificial intelligence) techniques in terms of locating services, how to asses them, put them together." That's very complicated, and again this is what Lisp is good at, Boetje says.

A couple of more rounds of seniors at the College of Charleston likely will have a chance to participate in creating the final product. Some of the things that still need to be worked on include a brand new compiler and implementing the common Lisp object system. Boetje hopes that by presenting the project at the conference he's also spurred interest from more parties in helping to identify what else the system needs, and perhaps may even find opportunities for his students to apply what they've done so far to some real-world problems raised by the IT community.