I’m under the impression that it must actually function in the contrary: i.e. exotic non von Neumann hardware requires new programming approaches.
Example: a systolic array, which you are only likely to encounter as an FPGA bitstream, by its very nature requires a different approach. Flow based programming is a better match for this specific system.
What we have seen since this paper was written is often the shoehorning of new concepts under control of standard von Neumann style programming, whatever the reasons may be, often including programmer familiarity.
Example: a systolic array, which you are only likely to encounter as an FPGA bitstream, by its very nature requires a different approach. Flow based programming is a better match for this specific system.
What we have seen since this paper was written is often the shoehorning of new concepts under control of standard von Neumann style programming, whatever the reasons may be, often including programmer familiarity.