|HOME||PUBLIC LIBRARY||ANOTHER PERSPECTIVE||INFOPERSPECTIVES||CONTACT|
They are number puzzles made of lesser but similar number puzzles. One kind is the spreadsheet (or mathematical model, if you prefer terms that give simple things undeserved dignity) sketching out the business of a conglomerate, of which IBM is a large and important example. Another kind is the increasingly popular recreation known as Sudoku. It's hard to say which example is more helpful, but for IBM's manager, investors, and customers, it's probably Sudoku, because there's a fair chance you can come up with a satisfactory result.
Sudoku is a puzzle most often in the form of a 9x9 numeric array made of a 3x3 array of 3x3 arrays. It starts with some of the squares filled in. To solve it, you put numbers in the remaining squares according to a simple rule. In the case of the 9x9, each smaller array must have one instance of the numbers 1 through 9. Every row and column in the larger array must also have one instance of each number 1 through 9. The number of possible puzzles is very large, so newspapers and web sites that run these puzzles can provide a fresh example every day.
The value of the numbers is irrelevant. What's important is that they are distinct symbols. So it's entirely okay to use letters or icons or colors to differentiate the values put in squares. Numbers just turn out to be the popular way to build these puzzles. Because Arabic numerals are understood the world over, they enable Sudoku to leap borders.
It's possible to create flawed Sudoku starting arrays, ones for which there is no solution or for which there is more than one solution. So, whether a Sudoku is created by a computer or a person, it should be checked for validity before it publication. This seems to have been the case so far, but enthusiasts might well know of flawed puzzles published, unintentionally, as valid ones.
The rules for the tables of numbers called business models are sometimes more complicated, but they have to follow rules, too, in their own ways. When the tables are called balance sheets, each half must at the end reach the same number; assets and liabilities must balance. When the tables are reports of revenue broken by category, the numbers have to correspond to details that arise within each business segment of a conglomerate, and segments have to correctly aggregate data from their divisions. The data may be altered to achieve a more pleasing solution in cases where managers and accountants have discretion. Often, a company has freedom to apportion costs like overhead to disparate operations as it sees fit, within broad limits.
When IBM publishes its numbers each quarter, interested observers sometimes look at the way various lines of business perform on their own and in relation to other lines of business. Some of these observers have particular concern for only one row or even one particular number in the IBM revenue table. Users, for instance, may focus on the number that shows results for the division that makes the equipment on which they depend. For instance, if you happen to be an iSeries shop, you probably want the iSeries component of the server row, which is broken out in yet another table by IBM, to show that your favorite product line has helped bring IBM shareholders prosperity. Weak number might be taken as a sign that IBM will be pressed to improve the iSeries product line, cut prices, or otherwise give it a little more pizzazz. On the other hand, they might be a signal that IBM seems to have lost the plot and that the iSeries could irreversibly lose some of its glory, the way the zSeries mainframe product lines has. Strong numbers make you feel like you're on the right track. Flat results, such as those reported for the iSeries in this year's first quarter, tell you that if you want clear answers, try Sudoku.
Outsiders cannot guess with great accuracy what the reaction inside IBM might be to good or bad results, and sometimes is seems that IBM's insiders won't have much better luck. So interpreting one quarter's results is pretty hard. But if some numbers in a specific piece of IBM are ugly quarter after quarter, both outsiders and insiders will guess that IBM might want to get out of the ailing segment. That's what happened in the case of IBM's PC business, even though it was a very large and, symbolically at least, a very important part of IBM's big picture. For quarter after quarter, those numbers were a sign, and not just any kind of sign; they were an exit sign.
IBM didn't drop out of the PC business precipitously, or without great forethought. But after due consideration, IBM actually started what became a major division half-heartedly, farming out product assembly, favoring an operating system from Microsoft, using chips from Intel and adopting some standards while, by virtue of the its enormous influence, creating many others. When the business took off in the early 1980s, IBM didn't bring manufacturing inside, make its own CPU chips, or develop a complete operating system of its own. (IBM's OS/2 used a lot of code licensed from Microsoft, keeping Big Blue on the hook.) So when it sold off its PC operations, it wasn't selling off a part of itself that was as deeply embedded in IBM as, say, the iSeries division.
Unfortunately, this will bring little comfort to users of iSeries machines or any other IBM servers who get worried when their division stumbles. IBM's 1991 sale of its typewriter and low end printer operations to Lexmark involved the disposal of a completely internal division that was nearly as old as IBM itself. IBM didn't let sentimentality stand in its way; it was completely focused on the numbers, like a Sudoku player working his newspaper puzzle on a commuter train.
Sudoku, like crossword, enjoys a natural marriage with newspapers. Both puzzles present the possibility of getting correct answers. That is something that does not seem to be offered by a paper's primary editorial content. News stories not only raise a lot of questions, but for the most part they provide compelling evidence that the experts, authorities, and powers that be don't have very good answers. But Sudoku didn't get its start in newspapers, at least not according to a sketch of the puzzle's history provided by Wikipedia. That account traces the roots of the puzzle to the late 1970s, when a puzzle magazine from New York first ran an example under the name Number Place.
|PCs & Printers||2,739||32||2,771||-17||-0.6|
|Systems & Technology||3,869||235||4,104||109||2.7|
|Total for All Segments||22,726||1,924||24,650||2,291||9.3|
|The Eliminations/other line reflects an adjustment for activities that involve sales among IBM divisions that later yield sales to customers.
In 1984, Sudoku appeared in a Japanese publication, Monthly Nikolist. It proved popular, and the form of the puzzle evolved. By 1986, Sudoku had blossomed into maturity. The convention of providing puzzles with no more than 30 squares filled in and the practice of setting out the pattern of seed numbers symmetrically had become features of the Sudoku culture. For two decades Sudoku grew in popularity in Japan, and was picked up by some dailies, but was still largely ignored elsewhere. Then, earlier this year, the puzzles began appearing in British fish wrappers, and from there has caught on in newspapers in the US and elsewhere.
It's probably just coincidence, but Sudoku's first appearance in America came at a time when Uncle Sam was having lots of trouble making his numbers work. In the wake of a huge jump in petroleum prices and some disturbing interruptions of supply, the US economy was struggling to adjust to a world in which monumental political mishaps could make things just too awful to get the kind of scrutiny they deserved. Americans, still sore from Vietnam, wanted a rest from trick questions and were eager for simpler questions that might have answers; Sudoku, working under its maiden name at the time, was a natural.
Japan in the second half of the 1980s was heading for its own rough ride, particularly in high technology. The immediate conditions gave some observers the notion that Japan would match or even beat the US in many aspects of computing, but Japan, unfortunately, was navigating by the wrong star: IBM.
In 1986, the year Sudoku popped up in Japan, IBM named its president John Akers, chairman of the board. Akers, like Presidents George Bush, was an Eli, and, more like the younger than the elder Bush, a specific kind of Eli that arguably jealous graduates of other Ivy League universities called a Coaster.
Akers had charm, style, sophistication, and charisma, but, it turned out, he had one small flaw that worked against him in Armonk: He didn't have any idea how to run IBM. Actually, his tragic flaw was much more specific than that. It came down to his apparent belief that colleagues and competitors were gentlemen, or would act like gentlemen. How this concept arose within IBM is easy enough to understand. IBM was enormously rich and powerful, so it could easily afford to develop its own mythology. How its brightest executives bought all that bunk is a little tougher to explain.
As it turned out, gentlemanly conduct was not abundant among a few of the people on which Akers and IBM depended, such as Bill Gates, Larry Ellison, all the Japanese computer makers, some if not most of IBM's division heads and every single one of IBM's largest customers. These folk were out for themselves and didn't see IBM as the deity Big Blue's lifers sometimes thought it was. It took IBM until 1993 to bring in Lou Gerstner, whose street fighting style and outsider's point of view saved the shareholders.
As IBM headed for a fall during the late 1980s, Japan, Inc. and the IBM it followed like a horse in serial harness, getting pretty much the same view of the world, were missing the salad days of the Internet, notwithstanding the role IBM systems at CERN had played in hosting the first World Wide Web, notwithstanding the enthusiasm with which Japanese culture embraced and enhanced every other advance in communications, from fax machines to mobile telephony. Corporate Japan also managed to overlook the high art to which manga had risen, a development that eventually helped a younger generation in Japan create web sites of immense allure, sites the West has yet to fully appreciate, and might not until it figures out that art seepage westward to China, South Asia, and the Middle East has become a daunting factor in cultural politics.
The reappearance of Sudoku in British papers also seems to be case of excellent timing. The UK is in a state of confusion about its government, its economic condition, and its culture. Basically, the Brits are getting worried about how they are going to make a living, even though they seem to be doing fairly well at the moment. This concern seems to be prevalent, with variations in kind and degree, in the other wealthy nations across Europe. But the British situation is compounded by its small but significant separation, geographically and politically, from the Continent. The UK has less to hang on to than its continental cousins. It's no wonder Sudoku went from a novelty to a craze in just weeks.
Even IBM's powerful and usually effective EMEA, which arose in part to keep the second generation of Watsons apart, and which has for ages dominated the market for central computer systems used by Europe's governments, banks, insurers and manufacturers, is going on a strict diet. It not just that IBM's growth in Europe has become problematic, it's that IBM's engine for growth, the services business, hasn't had the astounding success abroad that it often enjoys at home.
IBM's position isn't made any easier by the fact that the American computing companies most respected and admired across Europe are the ones that survived the dot com bubble to bring goods and services to the home, outfits like Google and Amazon and Ebay, not IBM, whose achievements are no longer so visible, nor Microsoft, whose visibility, if a count of headlines in the general press is any indication, often comes from its unfortunate role as the favorite target of script kiddies, phishers, and other denizens of computing's dark side.
Microsoft gets plenty of good press across Europe, too, but it's mainly about the company's gaming consoles, which turn out to involve IBM as a chip supplier. But Microsoft's game business may well be supported by the unwilling generosity of Office users and IBM's focus on chips for electronic toys probably contributed to its pending loss of the more prestigious opportunity to be the brains of the generation that will follow Apple's G5 desktop computers. (Apple's laptop computers use Power architecture chips made by Freescale, the offspring of Motorola, which run cooler than anything IBM can make.) As Apple makes its transition from Power to Intel processors, there will be plenty of interest in Sudoku among the company's investors, employees, software developers, and customers.
IBM's decision to let Apple move to Intel may or may not have been a choice based on the number puzzle that is the company's plan for its semiconductor manufacturing operations. So far, the reporters who follow the companies most carefully have not been able to ferret out a satisfactory inside story of the big switch. Still, however the decision was made there is going to be a big increase in Sudoku players at Microsoft. There's a lot more certainty in the puzzle than there is in trying to whether the operating system for Apple's Intel architecture machines will also run on PCs built to run Windows, boards that also support Linux, BSD, Solaris, and SCO. For now Apple is suggesting that its software will look for authentic Apple hardware, but nobody heard the word "forever" in any Apple statement.
If this particular puzzle doesn't fascinate you now, it might before too long, if Big Blue's bean counters play their spreadsheet variant of Sudoku and come to this conclusion: If IBM's services contracts that include workstation support could move client boxes from Windows to the safe and solid Apple OS XI (or whatever it will be called), a lot of client support charges built into the deals would fall right to the bottom line.
IBM would be in the chips because it got out of the chips.
— Hesh Wiener June 2005