My rule is: if I find myself sorting things into two buckets, one of which I’ve given a name that obviously means “good” and the other “bad,” that’s a sign I’m deciding emotionally rather than rationally. It’s time to take a step back and think about the distinctions between the things in the buckets harder.

Sorting code into “clean” and “dirty” buckets is a good example of this. Both bucket names are completely subjective, with “clean” obviously meaning good and “dirty” bad. As the article indicates, dig in a little deeper and it’s not hard to find objective ways the “dirty” code would actually be preferable.