In the 1990s the computing battlefield is the humble personal computer (PC): the war is for the right to control the multi-tasking disc operating system to drive it into the future. The reward is untold wealth for the winning manufacturer.
An operating system is the layer of computer instructions that sits between the hardware, the metal boxes and circuit boards inside a PC, and the software, the digital code which runs it. Applications, the software programs (word-processors, spreadsheets, databases, etc), ask the operating system for generalised services, for example, to open a file, display a character on the screen or read an input from the keyboard.
The original popular microcomputer system was developed in the 1970s by Gary Kildalls. Called CPM (Control Program for Microcomputers) it was a minimal system that drove simple computers based on Intel’s early 8080 processor. Kildall’s inexpensive CPM, once adapted to cope with the disc, screen and keyboard of each particular machine, provided a rallying point for software writers. Users of Apple II, one of the first personal computers on the market, needed a hardware adaptation containing an Intel processor to make their machine compatible with the growing band of CPM machines. This gadget was supplied by a teenage entrepreneur called Bill Gates.
Gates soon became the single biggest supplier of CPM. And it was to Gates’s company, Microsoft, that IBM turned in 1980 for an operating system to drive the next generation computer it planned to launch the following year. MS-DOS (Microsoft Disk Operating System) was born.
The problem with MS-DOS was that is demanded considerable computer knowledge of its users. Apple developed a different operating environment based on pictures (icons) and a hand-held pointer, called a mouse, which dispensed with the need for the complex command and codes. The graphical user interface (GUI) was created.
Gates’s answer to Apple’s product emerged in 1985 in the form of a product called Microsoft Windows. Windows put a GUI software layer on top of MS-DOS, transforming the machine into something resembling the user-friendly Macintosh.
Windows 3.0, launched in May 1990, attempted to bring the IBM PC platform on to a level playing field with the Apple Mac. Last year Microsoft was able to claim sales of well over 2m copies of Windows, making it the ”must-have” GUI for every MS-DOS user with a machine that would support it. In March 1992 Microsoft released Windows 3.1 which brought improved speed and stability and a neater, three-dimensional look to the product.
But rival software houses point out that Windows is really not much more than a mirage floating on top of the outdated and clumsy MS-DOS. Underneath every PC running a version of Windows 3.1 there is an operating system derived from the 1970s.
MS-DOS is not the only operating system belonging to a past age but ubiquitous today. Unix, a 20-year-old system originally designed for business computer systems, is standard on large systems. But unlike MS-DOS, Unix is predominantly an ”open” system which means individual companies can customise it to suit their own requirements.
The ”closed” nature of MS-DOS has brought Microsoft great financial rewards since, in effect, it controls the operating systems on 90% of PCs. But Unix vendors, themselves free to customise the basic Unix system, are eyeing Microsoft’s domination of the PC market. The latest challenge is called Destiny, from Unix Systems Laboratories (USL). One potent version, Unixware, comes with built-in connections to Novell’s Netware, the network operating system that already has 70% of the desktop market.
Unix is a ”multi-tasking” operating system, able to juggle more than one application at a time and run different services, such as printing, calculating and screen display, simultaneously, unlike the simpler MS-DOS. Windows, like the Macintosh System 7, offers a form of multi-tasking by inviting the applications to co-operate and pass the baton in an orderly fashion from one to another. But the systems have no central control, without which this kind of co-operative multi-tasking can fail spectacularly.
In 1987 IBM introduced a new multi-tasking operating system called OS2 to fulfil a longstanding promise to its 80286-based microcomputers customers.
Developed hand in hand with Microsoft, OS2 promised to be an early contender for the title of MS-DOS successor. In the event OS2 was a damp squib, badly designed (by IBM) and poorly implemented (by Microsoft). Microsoft departed from the alliance to renew its efforts on Windows. IBM doggedly picked over the ruins of OS2, and in the last two years, while Windows has become a spectacular market success, IBM has been quietly putting OS2 back together again as a powerful alternative that runs the same applications software as Windows, and more.
Against some 10m Windows installations, IBM can claim only 1m OS2 users, but the base is growing, soaking up potential Windows and Unix customers alike.
Positioned somewhere between the robust over-engineering of a full Unix system and the cheap and cheerful playground of Windows, OS2 was a healthy contender in the 1992 desktop operating system market. OS2 version 2.0, which trickled out a year ago, was a huge technical improvement, running Windows and MS-DOS software in multiple simultaneous sessions, and offering a clever graphical front end. Version 2.1, promised by the end of 1992, is the next page in the OS2 roadmap that leads, says IBM, to Taligent.
Taligent was the industry’s big surprise of last year. On October 2, 1991, IBM and Apple announced a strategic alliance designed to produce the desktop operating system of the future. Based on a long-standing internal Apple project called Pink, Taligent would combine the virtues of today’s IBM and Apple systems into an operating system of almost limitless power and unmatchable user-friendliness.
However, recent reports seeping out from the Taligent team speak of a research project that has grown uncontainably and will not see the light of day for years to come.
The most significant task for Taligent is to make a revolutionary new operating system compatible with its forerunners so that customers are not obliged to repurchase or reprogram their current software. ”IBM can’t launch a totally new operating system again,” says Joel Appelbaum, president and chief executive of the AT&TNovell alliance, Univel. ”Frankly, customers will kill ’em.”
Gates has been pondering the same problem. Windows was a tablecloth laid over the bare structure of MS-DOS that would attract users and software sellers into the world of the GUI without frightening anybody off. Gates knows that Windows is too shaky to contend in the longer term against Unix or even OS2 on the business desk top, so his next tricky task will be to take out the table without disturbing the tablecloth.
Enter Windows NT, a background project originally intended to be Microsoft’s own evolution of OS2, independent of IBM. To supervise its development Gates hired David Cutler, chief designer for VMS, DEC’s classic seventies operating system. Windows NT is currently circulating among software developers in an early experimental version, and looks identical to Windows, although Cutler’s more robust engine provides true multi-tasking.
Windows NT looks at first glance like a powerful contender if there is to be a clear winner in the desktop wars. But Microsoft’s confident presentations to corporate customers are being met by careful questioning. The issues they have to address are:
What is Microsoft’s track record in producing operating systems on its own?
None MS-DOS and OS2 were either originally developed outside Microsoft or were co-operative efforts with experienced outsiders, or both.
What is Microsoft’s track record in marketing multi-tasking operating systems?
Miserable the company flunked both Xenix and OS2.
Will Windows NT on the Intel platform be multiple-sourced by competing vendors like the various flavours of Unix?
No, it will be entirely proprietary to Microsoft.
Does Microsoft’s handling of DOS and Windows suggest that NT will quickly evolve to meet customers’ changing needs?
Hardly. The almost invisible progress of MS-DOS over a decade and the slow advance of Windows over seven years inspires no great confidence.
With the launch of Windows NT, the marketing of a new version of OS2 and the re-orientation of older Unix system to tap the desktop PC market, 1992 has undoubtedly been the year when the competition to control the PC operating system of the future opened out into full-blooded battle.