The University of Utah College of Engineering has established its sixth research center under Utah's Centers of Excellence program, a facility to develop new software systems for computers.
Dr. Robert R. Kessler, associate professor of computer science, will serve as director of the Center for Software Science, which received $100,000 in seed money from the state. The center also has funding support from Hewlett-Packard, IBM, and the Open Software Foundation, a consortium of major computer vendors.The College of Engineering now operates nearly one-fourth of the research centers created under the state-initiated Centers of Excellence Program.
Kessler formed the new center by merging the talents of the U. Computer Science Department's Portable Artificial Intelligence Support System project, known as PASS, with the departmental systems programming group. This relationship will result in an organization capable of tackling the most challenging research areas in systems software, he said.
PASS researchers are involved in designing computer systems that will use a wide range of artificial intelligence. The use of artificial intelligence simplifies the development of new software, whether applying artificial intelligence to practical problems, conducting related research, or with conventional programming techniques, he said.
Computer software generally falls into two areas. The first is system software that "operates" a computer, allows programmers to program the computer and enables users to run applications. Examples of system software are operating systems (DOS, OS/2, Unix, Xenix), programming languages and their compilers (BASIC, COBOL, FORTRAN, C, LISP), and programming environments (text editors, application generators).
The second area is that of application programs, which are created by a programmer using the system software to enable the computer to perform tasks such as spreadsheets, word-processing functions, desktop publishing capabilites and accounting procedures.
Scientists and engineers assigned to the center will concentrate on two areas of systems software: parallelism and operating, Kessler said. Parallelism is the use of multiple computers to work on a single problem simultaneously. Because many complex problems require vast memory and operating time, tasks can be divided and distributed to multiple computers that can solve these problems faster.
Initial parallel work will use the LISP programming language. LISP is an acronym for list processing. It is a high-level language primarily designed to process data represented as symbols instead of numbers. LISP is a popular language for prototyping software applications and for solving artificial intelligence problems, a field where computers are used to solve problems that are thought to require human intelligence.
"We've been working on LISP programming for years," Kessler said. "By adding parallelisms to LISP, researchers will be able to solve larger, more complex programs with greater efficiency."
Assistant director Jay Lepreau will direct the work on operating systems. "Our current effort is directed toward making the `mach' operating system the successor to the established Unix operating system," he said.
Unix is the standard base for sophisticated engineering applications. "Mach improves Unix with support for parallelism, better flexibility, and improved process communication. The microkernal approach of mach will also allow multiple operating systems to function simultaneously on a particular computer, giving users the ability to choose an operating system most appropriate for their problem.
"One way to think about the purpose of our research is that we are building software `tools' for others as well as for ourselves," Kessler said. "As we develop better tools, more researchers stand to benefit from the systems products. Our goal is to make software available to a lot more new users.
"During the first year or two, what we will attempt to do is take some of the research we've been doing, in which we have a great deal of expertise, and turn it into something from which people outside our research group and outside our department can benefit," he said.