I wonder if frameworks/libraries would count as such a system. For instance, consider the numpy library for Python (a library meant for fast, concise, and vectorized scientific computing). It has its own set of primitives (e.g., numpy arrays) and is optimized for fast vector computations. However, since it is written in Python/C, we still have all of the language features of those languages while using numpy. In this regard, it gives us many of the benefits without the cost.
I wonder if frameworks/libraries would count as such a system. For instance, consider the numpy library for Python (a library meant for fast, concise, and vectorized scientific computing). It has its own set of primitives (e.g., numpy arrays) and is optimized for fast vector computations. However, since it is written in Python/C, we still have all of the language features of those languages while using numpy. In this regard, it gives us many of the benefits without the cost.