My experience tells me that speed of defect resolution has only one primary factor and a bunch of far less important secondary factors. Secondary factors include quality of documentation, clarity of flow control, directness of access, developer correctness, and other factors.<p>The one thing that rules above all else is frequency of human iteration. If a person can test potential solutions 10 times a minute as opposed to waiting on a 10 minute code build then all the secondary factors become almost completely irrelevant. The higher speed of iteration allows the developer to learn and assess various different risk and performance factors that would be too expensive and timely to learn otherwise. That the developer can not only resolve the defect substantially faster but they can do so with less risks while simultaneously issuing performance improvements to the product. Code correctness becomes an irrelevant concern because through 100x greater iterations the given developer will learn from correctness violations and what that means for the given product.