Martin Banks, Personal Computer World 07/89 - checked

July 1989

Depending upon your point of view, I have either been sensible or very sinful for avoiding the rigours of becoming one of life's coders. Programming has never been my forte.

OK, so I can usually find my way around the vagaries of DOS command line programming, if programming it can be called, but I have never gotten my hands dirty with real coding. I have always been more interested in what has been happening with the technology in a broad sense, and the machinations of the industry.

You may now be wondering why I am indulging in a small bout of true confessions, so I will tell you. Because I have never been a coder, and in particular never been a personal computer coder of any variety, I have managed to steer clear of being trapped by the nature of the beast.

This does not mean that I feel any of the thousands (indeed millions) of happy hobbyist programmers are in any way 'trapped'. They can set themselves their intellectual objectives and have fun trying to reach them. When they succeed, they can have fun with (or even make a bob or two from) the result.

It is when they produce something that has the potential to make more than a bob or two that the trouble starts. To be fair, the chances of that happening are reducing. There are already a goodly number of successful companies selling most of the applications most of the people want to use most of the time, so there are fewer opportunities to find a money-making niche.

But there is another reason, I wonder if such programmers have the training, background or capabilities for such a job. In short, I'm not sure that they can hack it.

For the hobbyist programmer, some of the recent events in the world of PC programming may well be totally uninteresting. But for anyone who uses a PC for work, the increasingly apparent dilemma in which software companies now find themselves will be of great significance.

You don't need to have read too many issues of PCW to know that some of the major names in PC software, such as Louts (sorry, Lotus - why do I always make that keying error?), Ashton-Tate and Microsoft have all had one trait in common. They have all found it difficult, if not impossible, to produce new products on time.

They have all had different specific programming problems to overcome, and they have all found them more difficult to surmount than they obviously expected. That, in the end, may be the real problem that they face. I have begun to suspect that they do not necessarily fully comprehend the nature of the beast with which they are now dealing.

That lack of understanding stems, I believe, from the history of the individuals that make up the PC software business. Their background is, in many ways, the same as yours, dear reader. They too have sat in the darkness of the night, illumined by naught but the glow of their displays, hacking away at an idea for a program.

Yet when the likes of Microsoft's Bill Gates started that process, the personal computer was a very different thing. PC buffs sometimes play a double-think game, getting desperately excited about the potential power that the technology will put on their desks next week, next year, while forgetting that just a few years ago, the capabilities and facilities that Atari would now stuff in your pocket was beyond the wildest dreams of all but the mainframe user.

In those days, 16K bytes of memory and a cassette storage system were considered a big, posh personal computer. Disks, and disk operating systems, were rare until the de facto standardisation on CP/M, and 64K bytes of memory was still only 'theoretically possible'.

Programming in those days would, by definition, have been small and tidy, with few attempts at 'grand' applications. Even those that might be considered grand, such a databases or spreadsheets, were successful more because of their convenient (PC) packaging than the fullness of their capabilities.

Compare this with the situation today. 'PC' is now an inadequate terms for the machines available. They are, in most respects, the theoretical equivalents of mainframe systems. Yet, the programs which are used on them maintain their very direct lineage with past personal computer environments.

What is more, the companies producing those programs are trapped. They need to keep faith their existing users while trying to upgrade the products to make real use of the new capabilities the hardware platforms provide. Yet those products are the result of 'small'. To make them suitably 'big' has meant adding on bits of additional code to the original 'small' program. As this technique has continued over time applications have grown like topsy, often becoming extremely unwieldy and impractical objects in the process.

It is the continuation of that process that is now, I believe, causing the delays in the appearance of the latest versions of many old favourites. It may also be a contributor to the slow appearance of OS/2, which continues to come out in dribs and drabs from the portals of Microsoft.

Perhaps the most significant long term effect of all this, however, is that some of the currently established PC software companies may, in the fullness of time, disappear. They will go because the users will find a better source for their new applications. And that source will be the mainframe software companies.

If the PC is becoming the equivalent of a desktop mainframe, then who is better equipped, and has a better understanding of the environment, than such companies.

What is more, it is already happening. The mainframe companies are moving in, producing PC versions of their established applications. What is another more, they are selling.

All it needs now is for IBM to introduce a real desktop mainframe, one that runs one of its big-box operating systems, and the marketplace could move right away from the PC software companies for good.

Oh look! IBM is about to launch (in a very small way) the 7437, a desktop system that runs the VM operating system and is probably built round one of its special 370-architecture microprocessors. Most significant of all, the 7437 is a five-board add-in for a Model 60 or Model 80 PS/2.

I wonder if they'll do an EISA version?

end