My Write good papers paper gives a fairly reliable recipe to write good papers. I think it is difficult to good wrong if you follow this recipe.
How is it, then, that I can still write bad papers?
My experience is as follows: while I begin with a noble goal, the ideas become murkier over time. I tend to collect many small contributions, none of them large enough to constitute a solid research paper. The ideas aggregate into a mass, which makes up the bad paper. Each step I took is correct and a tiny bit interesting, but the sum fails to be compelling. Because it grows increasingly boring, I slip and my writing becomes sloppy.
I described it to my wife as follows: you tell a boring story, but some of the secondary characters are interesting. I am sure writers go through this process. I once read Dilbert‘s father say that when he started out, people did not like Dilbert. He was only interesting at the office. The solution? Throw away all of Dilbert’s life, except for the office part.
The solution? Throw away the paper and start from the most promising secondary contributions.
However, starting from scratch requires courage. How many software engineers throw away their code? How many companies throw away their product lines? How many writers throw away their novels?
Yet I am growing convinced that this trimming process makes the difference between the good and the great researchers. The people I respect and the people I admire. You have to be critical of your own work: throw away the worse of it.