The improvement to the Strip Path node is perfectly clear due to the node having a single input, and should "just be implemented".
The improvement to the Build Path node is more nuanced due to the node having two inputs. I agree with the suggestions made by Technico above, which address inputting a scalar and a 1D array, and vice versa. However, what should happen if the two inputs contain 1D arrays of different sizes? For example, what should happen if the first input (base path) contains 3 elements and the second input (name or rel path) contains 10 elements? How about if the array sizes were reversed (10 elements and 3 elements)?
For completeness, should the nodes also support 2D arrays (or even 3D and higher-dimension arrays)?