The 50% drop in the price of computing power every 18 months has permitted a patch and paste addition to existing software systems over the last 40 years. There was not a sufficient financial incentive to do a complete redesign. Hardware has never been more than 3% of the cost of operating a computer system. It is the software and operating it that cost 97%. This has resulted in most large organisations using mainframe software riddled with y2k bugs. For these systems, the documentation is incomplete or non existent. Finding all the offending code and fixing it by January 1, 2000 will not happen. The only alternative is to put in place wholly new hardware and software built to be y2k compliant. Remembering that 97% on the investment over 40 plus years has been in software and applications the correction needed to flush out this problem is enormous. The only question is how bad the dislocation will be. We are not of the doom and gloom school. We expect the problem to be solved but only after more dislocation and cost then currently expected or factored into financial projections.
The personal computer market has recently seen a shift in popularity to cheaper and less powerful computers. This is because, with the explosion of Internet, all that is needed for the average user is a computer that can communicate with the merchants on the net. They now provide undreamed of power and software capabilities for next to free. The so called network computer has already arrived. The error Ms. Reno and all of her ilk have made is to confuse objects with processes, things with ideas. There is no such thing as a personal computer or network computer as such. The designation is not dependant on anything that is possible to measure about the physical object or about what it can do. The designation is dependent on how it is used and how people now have started to use computers on the net has made the so called personal computer obsolete. The division of labour has changed. There is no need for the user to install or maintain new applications on a personal computer. The network is sufficiently reliable so that people can use a smaller and simpler computer connected to the net to accomplish far more then the so-called personal computer can today or the so-called mainframe did last year or the so-called supercomputer did a few years ago. Microsoft understands this. That is why it can not settle with the justice department if that requires it to stop bundling applications. It must continue to offer more and more applications for "free" or it will die. The statement many be too extreme. The most likely result will be a much lower share price: say $30. The Internet will make Microsoft irrelevant. This change in the process of how computing power and application software is used has resulted in the function of the operating system of personal computers being replaced by net browsers. There has been a dimensional shift. Ms. Reno is fighting over yesterday's non issue but for that error Microsoft's shareholders will pay.
The doubling of computer capacity every 18 months makes most computer technology products obsolete in very short order. Investors currently have put an enormous market premium on Microsoft is spite of the fact that it too is subject to technological obsolescence. Some time in the future investors will apply a discount. Timing, of course, is everything. The last large private investment banking house on Wall Street, Goldman Sachs, has decided to sell its' shares to the public. The insiders know when to sell to the widows and orphans. $30 billion of value will be created partly out of thin air. Each partner will get shares worth $125 million but only it John Q. Public is on the other side of the transaction buying what the insiders are selling. Insiders generally do better at market timing than the general public. We suggest it would be more prudent to buy shares of investment bankers when the insiders of those same investment banks are buying.