On Tuesday, 27 February 2018 at 19:01:57 UTC, jmh530 wrote:
The economic way of thinking is to consider whether the marginal benefit of a breaking change on all future code and whether that would exceed the marginal cost of a breaking change requiring old projects to be re-written. As most of us recognize, if the amount of old code that needs to be re-written is large and the cost of re-writing it is high, then it would overwhelm any changes of little benefit. Thus, I'm not sure this resistance to backward-incompatible changes is something all that specific to C++. I would guess that if D were as popular as C++, then the rational thing to do would be to be slow moving and be very careful about making costly breaking changes.
"The economic way of thinking is to consider whether the marginal benefit of a breaking change on all future code would exceed the marginal cost of a breaking change requiring old projects to be re-written."
