It is easy to notice that our cognitive abilities have much less problem processing a positive than a negative assertion. That is, it is just plain easier to understand that something is right, than saying something is not wrong, specially in long discourses like a mathematical demonstration, a lecture, thesis or paper.
I think this is the reasoning behind most lists on how to fail at something: instead of saying "don't these otherwise you you fail", one says that "do these in order to fail", often in a satirical mood. Interesting examples include a book called How to Fail: The Self-Hurt Guide, and this quick guide to fail in biology fields.
Scientific fields don't fall short of this trend, though. For example, an article was published last year entitled How Not to be a Bioinformatician. Rob Hyndman, in his blog (which I recently discovered looking for some LaTeX examples), has the post How to fail a PhD.
To contradict this trend, he also compiled the straightforward guide called How to avoid annoying a referee (which anybody in the scientific business should read and follow), expanding from this post in stats.stackexchange.com (which, in turn, follows the mindset described so far).
Of course one does not necessarily need to write guidelines in a jocular manner. One such (must-read) example is 1990's article from Gopen and Swan called The Science of Scientific Writing, in which they convey that most of the effort in communicating a result lies on the writer, and should not be deferred to the reader.
I find myself amused by the positive "how to fail" guides, though. They can make the point they want to address, while using a lighter tone. Maybe these are better ways of planting the seed of understanding something important, without taking the change of being perceived as boring.
