There are many reasons why a model numbering rather than a pure clock based nomenclature would make sense. First, you can consolidate several technical facts into one model name/number (FSB, cache, clock or additional features such as HyperThreading). Second, the confusion over different versions at the same clock speed could be finally reduced, making it simple for John Doe to figure out which model is the fastest. And third, there have been many examples in different industries that show how workable a model numbering is, such as AMD's nomenclature for the Opteron family, which is 14x, 24x and 84x. The first number represents the qualification for multi-processor usage: 1 is a single processor version, 2 for dual CPU machines etc. The x can be a 2, 4, 6 or 8, representing 1.6, 1.8, 2.0 or 2.2 GHz. That's it.

Last but not least, we should think about Intel's Pentium M series, which is going to grow with the 90 nm Dothan processor. As this chip will be considerably faster then Banias at reasonably increased clock speed, Intel would get into real trouble explaining why you would buy a 3 GHz desktop Prescott that could be outperformed by a 2.0 GHz Dothan core in certain applications.

According to our sources, Intel's clock speed number is to disappear completely in Intel's processor nomenclature. As the number of processors available is definitely not going to decrease, this is the only intelligent way of creating comprehensible product lines. Here is what the future of Intel processors will likely look like:

The Pentium 4 Processor will be called Pentium 4 Processor 5xx, while the Celeron family will be named Celeron 3xx.