关于code review 的一些见解
2007-07-13 17:31 by Jimmy.Lin
在之前的公司，曾经有段时间采用了code review. 但是发现方法好像都不是很好。所有的秘密大家请看如下。 这是今天从quest 那边下载的一篇文章里头的一段文字， 写的也很真实，就是我们曾经做过发生过的事情。为什么code review进行不下去的原因就是这个，但是我们的方法就是可以采用toad的TOAD’s CodeXpert
But even for those shops willing to spend the extra time and money, peer code reviews have another drawback that’s harder to quantify – team dynamics. Imagine that you’re a junior developer submitting your code for review, and the most senior guy on your team finds a really stupid mistake in your code. How would you feel as the junior guy? How would you feel as the senior guy? Without project management’s commitment and good team dynamics, the peer code review process can strain relations on many teams. I’ve seen people who’ve needed to iterate their code through the review process more than a few times – and both the author and the reviewers were stressed by the affair. Finally the project manager needs to keep the peer code review process and iteration counts separate from developer productivity for issues like annual salary reviews. I’ve seen a case where the most critical piece of application code took five iterations even though it was written by our most proficient PL/SQL developer. Complex logic can easily require more review iterations – regardless of the person authoring it and their skill level.
