Martin Banks, Personal Computer World 05/85 - checked

May 1985

If you look at the way the average company is organised, it's possible to state with a fair degree of theoretical certainty that there ought to be a good future for small multi-user computer systems. After all, the way they can provide small clusters of related workstations, coupled with the ability to network the clusters to build bigger systems, is a direct analogy of how most companies are departmentalised.

If this is the case, then it would seem reasonable at first glance to state that Unix is the operating system to go for and all else is dross.

However, it doesn't work out that way in practice. There are several different versions of Unix around, with even more Unix look-alike systems of varying degrees of compatibility. Although portability between machines is there in theory, in practice it just isn't that simple.

Despite this, it is still possible to suggest that Unix should become the dominant operating system for small to medium-sized business machines. Virtually any application that requires more than a solitary personal computer has to be a potential target for Unix-based applications.

One of the problems now facing the system is the publicity surrounding it. Having come from the rarefied air of AT&T's Bell Laboratories and the world's universities, it has been ill-equipped for the cut and thrust of the commercial world. In more practical terms, this has also meant that there have been few commercial applications packages available.

This might not have mattered too much it if hadn't been for IBM, which inconveniently introduced the PC just at the time when lots of micro manufacturers were looking closely at the potential market for supermicro machines running Unix.

The IBM machine, together with its host of clones and close compatibles, is the market for which the majority of the applications software has been written. Packages for Unix have, by comparison, tended to be fairly specialised and therefore expensive. Ironically, the world and its uncle has jumped on the IBM band-wagon; only Apple stands out as a major contender against the stream. Where Unix could have been a rallying point for all manufacturers wanting to get out from under IBM, it has suffered instead from everyone wanting to stay under the big blue umbrella.

The MS-DOS environment is now the de facto standard operating system for small micros - stand-alone machines at least. This still leaves some potential in the multi-user area, a potential that is not always satisfied by networking stand-alone machines together. They may appear to work well like that, but for many applications it will not be the best option.

Can Unix move in here? If it's going to then it has two years at most in which to achieve a worthwhile market-share. In that time, the fragmentation of the market for Unix systems must be cleared up. For example, although AT&T now claims that Unix System V is the standard, there is still a lot of System III around. At the same time, Microsoft's Xenix is claiming the lion's share of the low-end Unix market-place.

In this context it's not surprising then that Microsoft and AT&T have recently signed a deal to engineer a unified front, bringing Xenix and Unix into line with each other. Digital Research (DR), which recently completed a port of Unix onto the 80286 processor, has developed a library of 15 or more applications programs to run under the system. It has decided to drop the idea of marketing them, however, primarily because the fragmented state of the market makes the library's commercial viability suspect.

There will be some scope for more applications coming through as the 286 device becomes more prevalent. This may be helped by such things as language support for both Unix and the more prevalent IBM environment on the 286 processor. DR, for example, has a range of language products available for both the 286 and the Motorola 68000 that are source code-compatible with both Unix and its own Concurrent DOS. This, the company claims, will provide a funnel for software developers to go from one environment to the other.

But what is most likely to deal Unix its severest blow is the Apple Macintosh. This machine, with its fancy graphics, user interfacing, mouse et al has demonstrated that there is an alternative to the old character-based user interfaces typified by the dreaded A> prompt that we all know and love. Here is a machine that the non-expert can easily use because he can 'see' whet he wants to do from the icons shown on-screen, and point the cursor quickly and directly at the action required.

Couple this approach with the move towards high-resolution colour graphics, and you have a powerful tool for software developers to make their products really user-friendly for the first time. This is, however, a software technology that demands a considerable amount of local processing power to make it work properly. Unix, being dedicated to multi-user operations, will not always be available on machines that pack enough power to provide this for a large number of users, especially at the low end of the market where competition from machines with such capabilities will be found.

Such facilities are already available for the IBM PC environment. DR has already announced its Graphics Environment Manager (GEM), while Microsoft has Windows due this summer. GEM is interesting in multi-user terms as well: it will be part of DR's Concurrent DOS 286 for the 80286 device, which is itself a multi-user operating system.

The key to developments, as usual, will be what IBM does. With Top View it has shown a direct interest in the systems software market; it is unlikely to have missed the point of the Macintosh, GEM or the upcoming Windows. If it does produce something graphical by the end of the year, it will not only be the leader, but will completely crush the opposition. If it doesn't, then GEM or Windows could become market leader.

Either way, the place of Unix at the small systems end (especially with the emergence of Concurrent DOS 286 with GEM an integral part) looks to be increasingly insecure. Despite all the fine words said about it, everyone is hedging their bets like mad on the subject of Unix.

end