With Japan and West Germany the two apparent exponents of effortless success in the modern world, in its efforts to fix all that is wrong with its core business, IBM might send a large delegation of mainframe marketers to Japan so they can become inspired about engineering. In Germany, the same party, as we saw in CI No 1,399, might learn how to sow rather than harvest such things as customer goodwill. It is easiest to absorb information in an unfamiliar setting, one that poses the greatest possible threat: the unknown.

Aha!

At best, an IBM executive temporarily transplanted to a foreign clime might simultaneously get to appreciate the value of a new idea – like the importance of engineering or accumulation – and, by abstracting a bit, see that in both Japan and Germany there is a disadvantage to depending too much on one overly simple concept. We believe that a tour of client nations would so benefit IBM because the company is primed for change. There is a theory of learning that helps account for what we believe will soon happen – for IBM’s executives are indeed keeping a keen eye on the situations in all their important markets. One way to look at learning is to focus on the moment of insight, that instant when things fall into place. One absorbs facts and examples, but at first just files them away. Then, something happens. The retained information is assembled into a pattern. The pattern can be applied to new facts that are subsequently absorbed. As the body of factual knowledge grows, one modifies the pattern to include it all. When that millisecond of enlightenment occurs with sufficient force, it may provoke one to say one of the most beautiful words in the English language: Aha! It is more than remotely possible that IBM – the corporate culture, if not its individuals – has already reached the aha stage. But because quite a while might elapse between insight and the execution of a new plan, the company has wisely chosen to keep its trap shut. Customers are already a bit agitated. Actually, users need not spend a great deal of time speculating about the state of IBM’s enlightenment. Their strategies in the near term can be boiled down to two basic options. One is predicated on IBM’s having already figured out that it has played the marketing game out and is working feverishly to make the large systems its customers need. The other is based on the alternative postulate, the assumption that IBM is about one hundred percent ignorant, and won’t come across with a hot, new mainframe line. If IBM is on the right track, a customer would have to be nuts to invest heavily in today’s 3090s. –

By Hesh Wiener

The only sensible strategy is to get just enough MIPS to cover short-term growth and wait for the next generation. On the other hand – if IBM is simply unable to field better machines this year or next – the user’s strategy is also clear. The thing to do is get just enough MIPS to cover short-term growth and wait for the next generation. It all kind of puts IBM in a bind, doesn’t it? IBM seems to have gotten into this pickle by repeatedly optimising its behaviour for the short term. The company has depended on tactics when it should have thought out its strategy. At each stage of the 3090 game, IBM tried to do what would produce the best result during the next quarter, or maybe the rest of the year. This worked very well at first, even though the 3090 didn’t give customers a great deal more than the 308X line provided. In a way, this made life easy at big IBM shops, particularly during the years when the US economy was booming and nobody got too picky about computing budgets. Applications written years ago still ran, personnel didn’t have to learn unfamiliar techniques and managers could apply old rubrics to new problems. Unfortunately, computing technology moved forward apace, and not just hardware. Software has changed a great deal, too. But IBM opted to hold the line for as long as it could build machines using the basic engineering in the 308X and 3090 and sell the systems code in the MVS family. Because IBM i

s so big and resourceful, it simultaneously developed all the stuff that makes the 3090 line so inappropriate for these times. RISC architecture, relational data bases and a zillion other ideas that threaten IBM either originated or were perfected within its domain. The way it looks right now, if the death of the mainframe is imminent, the thing that will kill it is the kind of mainframe IBM is building and the way it has been oversold. The role of the personal computer – often and foolishly characterised as the organism that will destroy the large, central system – is a minor one in hardware terms. The main problem caused by micros is that they have put software development labs in the hands of millions of programmers, many of them amazingly talented. The software running on workstations uses a substantially different mix of resources than that written for mainframes. Because this software delivers so much that the user wants – and has made the micro compellingly attractive – an alert IBM would want to develop mainframes that can support multi-user versions of the best personal computer applications. This would entail coming up with a new mix of components that encourages users to buy or build programs that make large systems more friendly. The characteristics of a stunning mainframe have already been incorporated into several different types of mid-range systems, the simplest of which are the uniprocessors in workstations.

Happily pay

Basically, these machines deliver raw computing power on the cheap and support relatively inexpensive systems and applications programs. If the user wants to build up the more expensive parts of a machine – memory, fancy displays and file storage – there is an apparent and reasonable relationship between costs and results. This balance was once the hallmark of an IBM mainframe. No company had a more accurate sense of just what a customer would happily pay for and what seemed to cost more than it was worth. Last month, IBM raised the cost of its MIPS and software and cut the prices of main memory on its 3090s. While IBM’s marketing tactics may help the next quarter, they can only drive the progressive user away. This will prove as futile as the Japanese effort to engineer a permanently rising stock market or the German attempt to accumulate all Europe’s wealth.Copyright (C) 1990 Technology