Programming language holds key to future for Kodak subsidiary

Licensing, not product sales, boosts revenues

At a time of increasing prominence for software applications, Research Systems prefers to stay in the background. The company, a subsidiary of Eastman Kodak based in Boulder, CO, makes the software tools for developing these applications. In many ways, its IDL (interactive data language) is made to order for the imaging environment. The programming language is designed specifically to visualize very large, complex sets of data.

"Customers are looking for unique solutions to unique problems, so they need a language rather than a prepackaged application," said Jim Kelley, vice president of product marketing for the IDL business line. "They're typically dealing with large data sets and with data that are best understood when displayed visually."

Research Systems derives revenues primarily from license sales and ongoing software maintenance updates. The company also charges a modest runtime fee for the distribution of IDL programs or applications, he said.

The idea is to get programmers hooked on IDL so they write programs that practitioners want to use. Practitioners then sign up for licenses, or the application developers buy licenses to distribute to their customers.

License fees to use IDL now represent about half of the company's gross revenues, although Research Systems would not detail those revenues. Kelley expects income to grow considerably during the coming years, as the company expands its role of helping companies develop applications to meet their particular needs.

Until now, Research Systems has applied a hands-off approach to customers, concentrating on the programming language and leaving applications development alone. In a few cases, however, the company assigns its professional services group to help prototype an application.

IDL was the first product and is the cornerstone of the company, which was founded in 1977 by David Stern. The intent has always been to provide a powerful, flexible programming language that would enable researchers, scientists, and engineers, including those involved in medical imaging, to develop products and applications without the constraints of a prepackaged product.

Research Systems' first customer was in nuclear medicine. This discipline has continued to be a stronghold of the company, although medical imaging groups addressing other modalities have since signed on. Image processing, volumetric rendering, and tumor detection are some of the applications written with IDL.

Private companies, government agencies, and academia are major users of IDL. Longtime customer Siemens Medical Solutions uses IDL to process data, including the measurement of gated blood flow, in its e.soft nuclear medicine workstation. Neuroimaging researchers at Los Alamos National Laboratory are using IDL to visualize MR data for brain function studies.

Meanwhile, Research Systems continues to evolve its programming language. The latest version, IDL 5.6, was released in October. This product adds linear algebra routines designed to enhance the data analysis capability. It also has many new features for building IDL user interfaces, allowing even nonexpert programmers to easily create professional-looking user interfaces to their IDL applications.

With Research Systems' Watsyn software, released last year, the company can deliver a one-two punch for the medical imaging community. Watsyn, a cousin of IDL, is an implementation platform for scientists, researchers, and technical professionals who need help developing solutions for specific data analysis and visualization tasks. It allows application developers, including those involved in medical imaging, to reduce time to market for medical applications, maximize the value of their existing clinical workstations, and integrate applications into a common platform. Another key relative of IDL is ENVI, an image-processing product designed for airborne and satellite imagery.

"IDL will continue to evolve, serving the needs of the medical imaging, earth science, and science and engineering communities," Kelley said. "We're moving rapidly into multiprocessing and parallel computing. While computers are getting faster, the problems need to be solved even faster than the clocks on the computer. That's our impetus."