I think the pg's message is good on a higher level, though think about what would happen in an average software development team if everyone followed the rule.
Everyone would be working on fancy, interesting, big problems, while there would be no one to fix the leaking ceiling until after the sudden storm has started.
Oops, it seems it's actually the case in majority of the projects. Most people prefer doing new development than fixing bugs and writing docs, because it's way easier to crush hundreds of new lines of code than to write a good documentation, and it's more enjoyable than trying to reproduce a weird edge case. Also, usually new development is adding more business value than fixing bugs (provided that the whole product is not a one big bug).
Everyone would be working on fancy, interesting, big problems, while there would be no one to fix the leaking ceiling until after the sudden storm has started.
Oops, it seems it's actually the case in majority of the projects. Most people prefer doing new development than fixing bugs and writing docs, because it's way easier to crush hundreds of new lines of code than to write a good documentation, and it's more enjoyable than trying to reproduce a weird edge case. Also, usually new development is adding more business value than fixing bugs (provided that the whole product is not a one big bug).