At the heart of the modern personal computer lies a little piece of silicon we call the central processing unit (CPU) and it determines to a large extent what we can and cannot do.
In the early days of the desktop computer, the chip makers had to make a decision about how to make the CPU. One way to do it was to pack as much information as possible into each instruction that the CPU could perform.
The theory was that if you could do more at one time things would be faster (and faster is always good).
Another group of engineers said almost the exact opposite. They claimed the fastest results would be obtained if the instructions were extremely short but repeatable. (An instruction must be 'fetched' by the CPU and then processed. Fetching short instructions is obviously much faster than fetching long ones.) The first group became known as the complex instruction-set computing (CISC) and the second, reduced instruction-set computing (RISC).
Intel became the flag carrier for CISC while others such as IBM, Motorola, Sun, MIPS and Digital all went with RISC. RISC has proved to be considerably faster.
How is it that Intel, the world's leader in chip sales, could be so wrong? Well, in the first place, 'wrong' has many meanings. Intel has made more money than any of the others and the distinction is no longer quite so easy to make.
Sun Microsystems is the acknowledged leader of Unix work stations and servers and it is the only company left that produces everything: the CPU, the hardware and the software.