NT versus Unix is the title of a forthcoming report prepared for Unix International Inc with the help of Locus Computing Corp, Inglewood, California. But Chris Sanders of Locus claims that the versus may be overstating the situation. There’s probably a place in the marketplace for both, he says. While Unix International is currently keeping the report under wraps, Locus is keen to talk about its own conclusions. It’s hard to ignore the market dominance that Microsoft Corp brings to NT – it is the natural growth path for Windows users, who need a system to replace the MS-DOS spaghetti underneath Windows, claims Sanders. NT will sell in volume to users of personal productivity software, and take the lead in areas such as multimedia and workgroup computing, the report suggests. And some Microsoft-led protocols and specifications such as the MAPI Windows Messaging Interface and ODBC Open Data Base Connectivity are almost certain to be taken up as industry-wide standards, says Sanders. NT will be a sophisticated system, supporting multiprocessing and security at a fairly high-level.
Scalable and portable
As for Unix, that is a time-tested, used and abused system that is one of the most stable of the systems currently available he says. It is multi-user, supports sophisticated graphics via X Window, leads in networking options available, and is way ahead in distributed systems technology. It is also the most scalable and portable of operating systems. In its UnixWare implementation (from Univel Inc, the joint Novell Inc-Unix System Laboratories Inc operation, it will boast close integration with Novell NetWare and tight client-server integration with host Unix systems. The problem for strategists is that Microsoft is being somewhat vague about is own positioning of NT in the marketplace. Following the success of Windows 3.1, points out John Bondi of Locus, Microsoft has a mandatory requirement to build on the success of that product with NT. The NetWare-dominated local area network market will be the first sector to be addressed – exactly the same market at which UnixWare is being pitched. Windows for Workgroups, the first plank in Microsoft’s NT strategy, will provide a strong spur for current Windows users to move from an MS-DOS to an NT base, as they notice MS-DOS reaching its technical limits. But Sanders advises caution for those considering using NT for long-term mission-critical applications. It’s unproven technology, and Microsoft has not made it clear that it wants it as a server. Users, he says, should ask such questions as: does NT run on the system I want; do I have input into how NT is developed; and is Microsoft committed to open standards? Microsoft’s recent endorsement of the Open Software Foundation’s Distributed Computing Environment, the Object Management Group and the SQL Access Group (on whose work MAPI is based) may mark a change in attitude from Microsoft towards industry standards, but most of the underlying technology in Microsoft’s WOSA Windows Open Services Architecture is anything but open, and there are no guarantees that the application programming interfaces won’t change with successive releases. Other technologies – Object Linking and Embedding and Dynamic Data Exchange for instance – are well established, but are single-user-based and are going to have to be heavily modified to become network-aware. In theory, Microsoft holds all the aces, with everything from an enormous established base of Windows users to upgrade – but how easy will it be to move up? to the fact that it owns the thing lock, clock and barrel-shift, so provided it does play the game by the rules, there will be none of the problems of an endless variety of different and subtly incompatible Unix implementations to contend with. Which, again in theory, makes NT the most perfect open system imaginable – if, that is, it is written sufficiently well that applications really will be able to move between architectures with nothing worse than a quick recompilation. The 16-bit version of CP/M finally failed mainly because
although it looked more or less the same to the user on the 8086 or the 68000 as it did on the Z80, underneath it was quite different, and didn’t look at all the same to the applications. Positioning is indeed all-important – desktop, server or both, but Microsoft’s declared desire to see NT on every popular RISC suggests that in the medium term it is thinking in terms of it being a server as well as a desktop operating system, in which case those that believe they will one day see it on IBM Corp’s ES/9000 mainframes and reckon that it will prove much more attractive an option than MVS when it gets there, could be proved right. But if the thing really does live up to its potential and become a runaway success, a very uncomfortable conflict of interest will arise: users will love the fact that if they find that Digital Equipment Corp is not, in their view, a nice company to do business with, next time around they will be able to call in Hewlett-Packard Co without a moment’s worry, safe in the knowledge that their workload will transfer across in a weekend – but the likes of Novell, Borland International Inc and Lotus Development Corp will find the fact that Microsoft increasingly owns the entire software industry intolerable. And the powers that be are likely to agree with them: if a monopoly is big and valuable enough, almost any action can be construed as abuse of it, and Microsoft will be punished unmercifully for its own success.
Tiny handful
It will likely be required to hand NT over to some kind of trust if it wants to remain in the spreadsheet, database and languages businesses – however much users clamour that they want Microsoft to remain the way it is. Anyone that doubts that that is the line Uncle Sam will take need only look at the measures taken against importers of low-cost chips, which penalised a vast number of computer manufacturers and buyers to protect a tiny handful of US chipmakers, and against Japanese active matrix flat-panel display makers, which had the impressively detrimental effect of forcing all US vendors of notebook computers with high-resolution screens to make the things abroad, while doing nothing for America’s own vestigial falt-panel display-making industry. And where will all the operating environments coming down the pike, like Pink and Sun Microsystems Inc’s Spring overlay for Solaris Unix fit into all this? Answer came there none. The Unix International report, when it emerges, will raise some important questions for those wavering between UnixWare and NT. But the battleground is still being marked out, and Unix is likely to find as many problems and hurdles reaching the lower end of the market as Microsoft with Windows NT will encounter moving up.
