WHO CAN REALLY SEE THE FUTURE?

  Computer industry analysts have predicted the revolution of computing for years, they are often wrong. Predictions such as: there will only need to be four computers in the entire United States because of their huge operating cost and power; that personal computers will mean the whole sale replacement of large scale mainframe based computing; that the Client Server Computing model will reduce the cost and the complexity of computing; that IBM and Microsoft's OS/2 operating system will be the dominate desktop operating system in the nineties. These statements have been the reason for many unscheduled career moves, usually by the people taking rather than those giving advice.

The reality is that no major computer system or environment has been replaced by the newer models. Mainframe computers are being sold and used more today than they have been in the past. Mini-computers have not replaced Mainframes nor have they been replaced by high end personal computers. Client Server Computing has reduced cost in some area and greatly increased cost in others. The industry is embracing open standards such as TCP/IP, HTTP, LDAP, X.500 , POP3, SNMP, SMTP/MINE while at the same time buying propriety single vendor solutions. How can the contradictory nature of our industry be explained? How can industry analyst who make wild inaccurate predictions continue to hold favour? The answer is because by our very nature, the computer industry is evolutionary not revolutionary.

The working computer professionals must deliver solutions to their customers, unlike the working press whose job it is to report on change. Evolutionary predictions have a much greater chance of being proven correct than revolutionary predications however, revolutionary predications are sexier and sell a lot more advertising space. Industry analyst can also rephrase, rewrite or even take credit for the new realities that ruin careers in other industries. Senior management, in most companies, do not truly understand computer related issues and rightfully rely on their Information System people to make a case for the huge investments in technologies. This investment is necessary to stay competitive in the market place. These professionals can use reports to substantiate almost any view they want. Opposing views and recommendations are common because they are often underwritten by the computer vendors themselves. In other industries, analysts would be removed for reports that are overly influenced by vendors, inaccurate or just plain dumb. Let us look at the automotive industry. If in the 70's automotive analyst urged their readers to produce larger cars and ignore smaller cars from Japan because the Japanese failed to grasp the American market. That analyst would have a good chance at selling cars but at a much more personal level like at a used car lot. If in the 90's an analyst reported that smaller cars should be produced and the industry should stay away from producing mini-vans, sport utility vehicles, and high performance luxury cars he would also have a chance to sell at a much more personal level. Computer analyst seem to be able to report that they influenced these changes. Analysts have reported: That Windows 1.0 was a failure and that Windows will never succeed in the marketplace; that Oracle, 3Com, HP, and Compaq will never recover from their problems; that OS/2 will be the predominate desktop operating system in the 90's; that the Client Server model will reduce cost and replace host based systems. These same analyst now take credit for: improving Windows; getting Oracle, 3Com, HP, and Compaq back on track; and that helped Microsoft evolve OS/2 into Windows NT. In other words, Automotive analyst that are wrong become Automotive salesmen. Computer analyst are just never wrong.

Long term predictions must be general in nature if they stand any chance at being accurate. These predictions include: that computers will become faster, more powerful and easier to use; that telecommunication bandwidth will become cheaper; that content and information will become more important for businesses to stay competitive. These may seem like unimportant predications, but can be used to determine the validity of shorter term predications. The debate about computer languages should be viewed using the predication about faster computers. Analyst reported that C and C++ could never replace assembler level coding because C and C++ programs will be too slow. This would be the case if we tried to write Server, Database and CAD/CAM applications to the 8088 processor. Java and ActiveX applications could also not thrive if processing power does not increase, but processor speed will increase and cross platform modular application development will become the dominate form of all application development.

Short term predictions are often not predictions at all, they are just acceptance of the facts. These predictions include: the network will become the computer; that Web browsers allow the Internet to reach an ever growing customer base; that Web browsers are losing their importance as they become a commodity. Some of these predictions are just marketing hype. Customers moved from DOS 2.0 to DOS 3.0 to DOS 5.0, and Windows 3.0 to Windows 3.1, to Windows 95 and Windows NT because they improved. They where evolutionary not revolutionary they didn't change the model they just basically crashed less often. The prediction that customers are going to move to the current release of the OS is not a visionary statement especially since the new OS is bundled with all newly purchased Intel hardware.

Mid-term predications are the most interesting. They require some vision and cannot rely exclusively on market hype. These predictions are especially fun now that the industry is moving so quickly. I am most intrigued when a CIO hands me a 10 year plan that covers what systems his company will be using. It is like the predictions that Man will walk on the moon made in the 1700's, they where right, except the astronauts did not get there by being shot out of a cannon and did not wear three piece suits on landing.

In the next five years there will be more, not less, operating systems. Many will come from Microsoft and be a direct descendent of Windows. Unix and legacy operating systems will maintain an excellent market share. New special purpose operating systems and environments will find a huge market; these will be in routers, high speed caching system, network computers, set-top boxes. These operating systems will allow for highly specialized high performance system. They will integrate into your existing environments through cross platform applications . Much of the overhead associated with today's operating system will be striped out of these systems. For example: these operating system will allow data-based servers that make low level system calls; there will be specialized system that allow very good 3D modelling for virtual reality applications; these modelling system will interact with the information stored on the data-base servers providing cross system interaction for simulations; applications that need to run on multiple systems will be written in languages such as Java allowing them to run where the data resides or where the infrastructure of the network requires.

A large portion of today's operation systems are really utilities that allow the user to perform maintenance functions like copy files, change user preferences, load and control applications. Data-Base servers do not need an active graphical interface. High speed caching file severs do not need complex virtual memory systems. These services can be loaded, run and most importantly unload dynamically by the network. As hardware cost continue to drop dedicated single purpose devices make more sense. Cross platform applications can provide the ease of use and integration with network management systems and directory services for the seamless manageability of these devices. Many of these devices will be managed across high speed data networks by service providers. Marc Andreessen, Co-Fonder of Netscape, has talked about devices that will access information that are free to the consumer just as many customers do not directly pay for cellular phones. The service provider can afford to deliver a specialized low cost hardware device so that customers can access their service or product: a specialized on-demand video box from Blockbuster; a game console from the Sega Channel; a data base server from your data warehouse provider; a manageable home security device from your security company; an application runtime environment from the company that wants to lease you office productivity software. The industry understands the value of open systems. As it is forced to deliver on that promise, businesses will be allowed to capitalize on innovation like never before. A single box or system does not make economic sense. It ties your future to your suppliers future. If they deliver a substandard solution at a higher cost than their competitors than your competitors benefit. The integration of high performance, large scale, highly reliable and distributed systems that delivers a lower cost solution will make companies successful in the next century. Specialized operating systems and dedicated single function hardware will free businesses and consumers to make real choices.

Businesses that reject this new model will be force to relive the IBM model. Lower cost computing power did not move customers away from the single vendor solution, VisiCalc did. That is to say that innovation was restrained by the model. VisiCalc allowed accounts to make one change to a number, and have the computer recalculate the whole spreadsheet, seems like a simple idea. But the IBM model did not allow it more importantly they did not seem to care that it served a real end-user need. Vendors like IBM, AT& T, Microsoft, Oracle, Sun, Novell, Netscape and many others are working on building systems that can interoperate. These systems can be centrally managed by an internal staff or externally managed by professional service organizations. The ability to choose the right systems and applications, tying them into your existing legacy applications will allow Information System professionals to provide the lower cost of a centrally managed system and the innovation delivered on personal computer platforms.