Protecting hardware devices from unwanted software attacks is a current area of major security concern. Coupled with the need to secure and verify data sent to and from such devices, the need to supply systems capable of uniquely identifying and securing hardware devices is considerable and imminent. This paper introduces techniques which possess the potential to generate unique identifying codes for given hardware devices based on measurable quantities or features associated with the given hardware and software configurations executing upon it. The techniques are investigated by considering abstract properties in order to validate primarily the feature normalization techniques employed prior to the code generation phase which allows features with highly variable distributions, and whose component values are independent of each other, to be employed within the code generation system.