我有一个同事，他坚持认为他的代码不需要注释，这是“自文档”。
我已经审阅了他的代码，虽然它比我看到的其他人编写的代码更清晰，但我仍然不同意自文档化代码与经过注释和文档化的代码一样完整和有用。
帮我理解一下他的观点。
什么是自文档代码 它真的能取代注释良好和文档化的代码吗 在某些情况下，它是否比有良好文档和注释的代码更好 是否存在代码不可能在没有注释的情况下自文档化的例子
也许这只是我自身的局限性，但我看不出这怎么能成为一种好的练习。
这并不是一个争论——请不要提出为什么注释良好并有文档记录的代码是高优先级的原因——有很多资源都表明了这一点，但它们对我的同行来说并没有说服力。我认为我需要更全面地了解他的观点，才能说服他。如果你有必要，可以提出一个新的问题，但不要在这里争论。
另外，那些反对自我记录代码的人——这主要是为了帮助我理解自我记录代码传播者的观点(即积极的方面)。
当前回答
我曾经和一个家伙一起工作，他打算把金融套件卖给一家大公司。他们坚持让他记录源代码，他写了一个30多页的汇编程序，并说“这是有记录的，看”——然后他翻到第13页，有一条评论“bump counter by one”。 伟大的产品，伟大的实现者，但是……
无论如何，在我看来，上面的重要评论是为了设置上下文。这段代码是自记录的:
> from BeautifulSoup import
> BeautifulSoup, Tag def
> replace_a_href_with_span(soup):
> links = soup.findAll("a")
> for link in links:
> tag = Tag(soup, "span", [("class", "looksLikeLink")])
> tag.contents = link.contents
> link.replaceWith(tag)
但是，就我个人而言，需要一个背景来充分理解它。
其他回答
我会扭转局面。
问问自己在他的代码中有什么不理解的，然后让他把这些记录下来。也许你也可以告诉我们一些。
代码本身总是对代码功能的最新解释，但在我看来，它很难解释意图，这是注释最重要的方面。如果代码写得很好，我们已经知道代码的功能，我们只需要知道它到底为什么这样做!
不管纯粹的自文档代码是否可以实现，有一些事情是人们应该做的:
Never have code that is "surprising". Ie. don't use silly macro's to redefine things etc. Don't misuse operator overloading, don't try to be smart on this. Split away code at the right point. Use proper abstractions. Instead of inlining a rolling buffer (a buffer with fixed length, with two pointers that gets items added at one end and removed at the other), use an abstraction with a proper name. Keep function complexity low. If it gets too long or complex, try to split it out into other other functions.
当实现特定的复杂算法时，添加描述算法的文档(或链接)。但在这种情况下，要努力去除不必要的复杂性，增加易读性，因为很容易犯错误。
This is an excellent question. It traces back to the first programming language that allowed comments, I'm sure. The code certainly should be as self-documenting as possible. Comments that point out the obvious, should be eliminated. Comments that make it easier to understand the intent, purpose, and use of a given method or section of code can be invaluable to those of us dolts that may be less familiar with the language or code in question. Structured comments that allow for the generation of API documentation are a good example. Just don't comment an IF statement that checks to see if a checkbox is checked and tell me that you're checking to see if the checkbox is checked. Restating the obvious in a comment is the worst waste keystrokes in our universe.
//For example, the above text deals with what is a useful comment
Self documenting code is code that is trivially easy to understand. Variable naming goes a long way to making code self documenting, but i find the best tactic is to break any complicated logic down into tiny little chunks and refactor that information into seperate methods with verbose and informative names. Then your complicated methods become simply a list of steps to be performed. The tiny private helper methods then are documented sufficiently by their own method name and the complicated methods are documented as a sequence of abstract steps to be performed. In practice this strategy cannot always be applied perfectly so comments are still very useful. Plus you should never completely abandon any tool which will help you write code that is easier to understand.
