Don't Forget the Developer
Another strategic target in the operating-system competition is the applications developers in large corporations. These developers are the focus of two conflicting pressures: downsizing (and the scattering of corporate information system resources) and the increasing need to develop applications faster. The beleaguered developer presents a critical support issue and a strategic opportunity to operating-system vendors. The operating system that best supports internal developers has a real edge in getting their company's business.
"One of the biggest problems you have in corporate America is getting an application developed," observes Mike Colleary, IBM's marketing manager for OS/2. "Corporations need applications that give them a competitive advantage, something that everybody and his brother can't buy off the shelf." According to Colleary, up to 2 million people worldwide are developing unique applications within companies. Naturally, he sees OS/2's strength as a development platform.
Support Is Critical
Once an operating-system vendor has set its sights on mission-critical applications and corporate information systems, it must address a number of complex issues - issues that historically have had little to do with technology. This implies a higher level of support for the customer than simply delivering a shrink-wrapped software package. The catch is that while corporate customers may demand premium levels of support, they also demand low prices. The challenge for the operating-system developer is to define a collection of capabilities that can reduce the cost of support without overburdening the operating system.
IBM's Colleary claims that that profile of the user is the model on which to build a support infrastructure. He says the operating system should be as transparent to the user as possible and offer accessible help functions.
This is a new paradigm. Corporate customers are accustomed to having fast, effective, and comprehensive service and support from mainframe and minicomputer vendors. When something goes wrong with mission-critical applications, they expect a fast solution.
Arun Taneja, vice president of marketing for Univel, predicts that "the guy who is going to provide that cradle-to-grave service is the one that is going to win in terms of satisfying the MIS community and satisfying the Fortune 1000 community." Univel is targeting the systems integrators and system hardware OEMs as well. Taneja says that all the largest systems integrators are working with NetWare and are familiar with Unix, a situation that greatly improves Univel's chances of acceptance for its desktop operating system, UnixWare.
Corporate Change
Most corporate customers are no longer dazzled by the technology of desktop systems. They've traveled the path of the PC evolution . These customers are no longer looking for raw technology; they want efficient tools. Corporate customers expect their tools to work consistently and at the least possible cost in time, effort, and money.
The current wave of what is variously referred to as reengineering, rightsizing, or downsizing has one immediate goal: to reduce the cost of doing business. Downsizing directly affects the development of operating systems by the changes it makes in both the information systems infrastructure and the expectations of the people who own and use those systems. The other key motive behind corporate downsizing, heard most often from those users who prefer the term reengineering, is the flexibility that it offers.
The infrastructure of corporate information systems, at least for the purposes of this article, can be viewed from three perspectives -- what it does, who supports it, and who pays for it. The move in recent years from mainframebased information systems to those based on minicomputers and microcomputers was a user revolution driven by cost (primarily of the computing hardware), freedom (of information access and exchange), and focus (the mainframe-based information system focused on corporate-level problems, often ignoring department-level problems).