We really need to stop measuring "compatibility" numerically. Its dishonest and encourages engineering decisions to be made instead of policy ones, when the dat

We really need to stop measuring "compatibility" numerically. Its dishonest and encourages engineering decisions to be made instead of policy ones, when the data is purely the result of someone else's policy decision. Not all things have equal weight; choosing which things to call an element, which ones to merge, which ones to split is always subjective. Valid discussions are where we fully state our chosen assumptions, not ones where we then make fake data out of them.