I don't agree and I don't understand their dichotomy. That whole paragraph is very confusing and lacks context. Did they have good test coverage from the get g

I don't agree and I don't understand their dichotomy. That whole paragraph is very confusing and lacks context. Did they have good test coverage from the get go or not? If they did, why was so much support caused by failing code? If not, it's easy in hindsight to say this but there's no guarantee your tests would have caught all such support issues.