It Will Never Work in Theory

We welcome critical discussion of the papers discussed here, and of our comments on them, but abusive language of any form will be removed. Please also ask yourself the following questions when posting:

1. Would you express your opinion so strongly if the paper had come to the opposite conclusion? For example, if it had found that doing X actually did improve code quality, or that Y wasn't better than Z, would you be as upset as you are?

2. Are you criticizing the paper's actual claims? If it says that P holds for novices who are learning how to program, there is no point saying, "But no experienced programmer would do that!" Similarly, if the paper compares examples found on the web that were created using tool F with equivalent examples created using the authors' new tool G, the fact that those examples aren't on the web is irrelevant.

3. Are your criticisms of the paper's statistics numerical? It's meaningless to say, "The paper's sample size is too small," without some quantification.

4. Are you employing proof by rhetoric? The statement, "It's obvious that J," is verbal bullying, not proof. And many "obvious" things aren't actually true: that's why we do studies.

Finally, we strongly prefer commenters to identify themselves, as it encourages civility and moderation. We are therefore much more likely to let strongly-worded comments stand if their authors are willing to stand behind their words, just as the authors of the papers we present are required to stand behind theirs.

Comments powered by Disqus