If I inject a new feature into production and it destabilizes existing functionality, it’s the new features fault (and it gets rolled back). It doesn’t matter i

If I inject a new feature into production and it destabilizes existing functionality, it’s the new features fault (and it gets rolled back). It doesn’t matter if I really think it’s the fault of the other component, or I would have designed it differently (especially now that I know the consequences!). Maybe we end up “fixing” the existing functionality and then proceeding with the new feature — but not until stable state is re-established. This is like standard practice for shit that is of trivial importance to humanity in a newish engineering field. So I don’t for the life of me understand the logic of being more flippant when it comes to human lives in a more established engineering field.