Design is the process of producing an implementation ready to be laid out into a chip, onto a board or a combination of both. What happens before this point is called functional design and the functions performed after that physical implementation. The argument is that the design process should be independent of the technology being used for its implementation, and while that held true in the past, today the line has become quite broad and the issues traversing the line continue to grow. However, it is still useful to think of it being a two stage process so long as a simple over-the-wall transfer is not implied.
For the digital portions of the design, the design is normally transferred as a register transfer level (RTL) description that defines the circuitry in terms clocks, registers and the logic that exists between them and been shown to be implementable in terms of timing and other considerations that are influenced by the back end. In some cases, a gate level representation is necessary.
The RTL code may be the input description, or as is becoming more common, a higher level of abstraction may be used for the initial description. In addition, increasing amounts of the design are brought in through the selection of semiconductor intellectual property (SIP), or normally just referred to as IP.
The analog portions of the design are not amenable to this kind of separation since the devices are much more influenced by the physical aspects of the implementation. While the analog and digital portions may be logically integrated for the purposes of verification, the implementation of both parts happens independently apart from final integration at the chip level and at points in the flow, such as floor planning.