TFH Flashback: Brave Two Worlds, November 1992

In the last contest between IBM and Microsoft–the battle for control of personal computers–Microsoft won. OS/2 is growing, to be sure. But nearly all the personal computers going into business offices use applications written for a Microsoft environment: either DOS or Windows. With Windows for Workgroups now reaching the first brave users, Microsoft’s main systems software development effort is aimed at controlling the fastest personal computers in every business. This is the project to field Windows NT, a software product already well along the way to a mid-1993 debut.

In addition to running on fast personal computers, Windows NT is expected to run on several important brands of midrange systems. In particular, it will work with machines made by DEC and Hewlett-Packard, the two makers of midrange computers that are in IBM’s league when it comes to wealth, power, reliability and technical prowess.

Within a year or so after it hits the market, the Windows NT program will most likely vanquish Unix, at least in commercial settings. It will probably make it easy for businesses, professional firms and corporate departments to operate networks of PCs. Even if it doesn’t make running a network easy, it will almost certainly bring it within the scope of many businesses that do not have computer experts on the payroll. A further advantage of Windows NT–at least in Microsoft’s battle plans, as we understand them–will be its standard data base management system. This program, called Microsoft Access, is the key to NT’s success beyond the personal computer world. If Microsoft can make it good and keep it cheap, it will be the strongest selling point of the NT environment. By the end of 1993 or during 1994, barring tragedy at Microsoft, NT will be the only practical software alternative to OS/400 in midrange computing. And even though no single hardware company that licenses NT will be as big as IBM–or even as big as IBM’s AS/400 division–the aggregate could well be larger, richer, stronger and faster on its feet.

IBM and, more importantly, AS/400 customers, had better keep an eye on NT starting now. AS/400 customers have two responsibilities. First, keep IBM very frightened so that it continues to deliver improved technology and better value. Second, develop a complete and accurate understanding of NT so that political challenges to an AS/400 strategy within the corporation can be thwarted.

Although there will be many settings in which NT and a network is superior to OS/400 and an AS/400, if IBM does a good job it should prevail. Under no circumstances should IBM allow itself to fall behind. Microsoft is to today’s computer industry what IBM was in the past and what it could be again: A company that gets to the head of the pack and stays there.

Unix, a favorite in technical circles and currently an increasingly popular alternative in commercial set- tings, won’t disappear. But it will be a distant third for bread-and-butter business computing. Unix buffs in academic as well as industrial settings will remain the most creative software developers and the companies building platforms for the environment will also retain a lofty stature. But the ordinary business trying to keep its books in order, provide support for sales and service personnel and generally avoid harboring a cadre of computer experts who cannot speak plain English will shun Unix. The chores done by AS/400s and NT platforms will be taxing (and, often, mystifying) enough to non-technical personnel. Unix will be over the top.

Next year, when the opposition of NT and OS/400 becomes more obvious–to IBM, too, we hope–the competitive spirit of IBM will rise. Extensions to OS/400 that provide better support for personal computers will get the attention of IBM’s software developers. Alternatively, IBM may field a line of terminals that have the power of personal computers (and maybe a special version of OS/2) but the look and feel of an extended 5250-type tube. In either case, there will be a need for desktop applications that mesh with programs on the AS/400 host. If IBM cannot create the terminals by itself, it might get them as a result of its collaborative efforts with Apple.

For all we know, IBM might even buy Apple, or at least all the rights to the desktop computers and software the joint venture is producing. The partnership, as it now exists, may merely be IBM’s way of obtaining access to the skills within Apple while avoiding the kinds of problems the IBM corporate atmosphere can cause in create organizations. If this is the case, it shows that IBM learned a great deal from its unfortunate experience with ROLM, the once-fruitful telephone equipment company that went to seed after IBM acquired control.

Our hope for the AS/400 is in no way meant to suggest that Microsoft is a pushover. On the contrary. Microsoft is precisely the kind of rival IBM’s AS/400 group needs to bring it to sharp focus. And the autonomy that the AS/400 group is gaining is absolutely necessary . . . not only for the success of the AS/400, but for the future of IBM, too.

The OS/400 and Windows NT worlds will each be better as a direct consequence of the struggle. And by better we mean better for the customer.