A processor's configuration is fairly straightforward. There is a clock speed at which the system runs. The Intel Pentium 4 Prescott clock speed is based on its Front Side Bus (FSB), which currently is clocked at 200 MHz and runs at four data transfers per clock cycle (quad-pumped or quad data rate), which is where the marketing-friendly term FSB800 came from.
The actual processor clock speed is derived by applying a multiplier. The Pentium 4 Processor 540 at 3.2 GHz runs at a 200 MHz FSB speed and a multiplier of x16, so this is very simple math.
A couple of years ago, you could easily overclock a processor by simply choosing a higher multiplier. A Pentium II 266 MHz at 66 MHz FSB speed could easily be operated at 300 MHz core clock speed by changing the multiplier from x4.0 to x4.5. There were only few models that caused problems from a technical point of view then.
Obviously Intel was aware of the fact that some buyers would restrain from buying the fastest processors, as there was a high probability that a mid-range model would run equally fast. As a result, the chipmaker decided to factory-lock the multiplier to match the clock speed for which it was validated and labeled.
A second way to overclock involves the FSB clock speed. This way of overclocking has by far more impact on performance, as you do not only drive the processor faster, but all the other chipset components such as the memory controller, storage interface and AGP/PCI devices will be comparably overclocked. Nowadays, chipset and motherboard makers decouple the AGP/PCI clock from the FSB clock in order to not have overclocking attempts fail by clock-sensitive graphics or add-on cards.