PROGRAMMING : Facts and Fallacies of Software Development by Robert Glass

You know, I dissed this book a bit in my review, but a strange thing has happened. Whenever someone at work complains about a particular 'bad smell' about their project I think to myself 'oh yes, that's such and such a fact' from the book. It seems that it is useful for collecting the well known ideas we have and giving them namees, a bit like the pattern books.
I think I might be a total nerd and have a T-Shirt printed with
25% increase in problem complexity => 100% increase in solution complexity
printed on it...

Oh, that reminds me. I have been a bit of a fan of the 80/20 rule (80% of a problem is solved in 20% of the time, or 20% of the problem takes 80% of the time) After doing a little analysis I think it's a bit excessive; if you assume that it is a 'linear' (actually geometric) relationship, and apply the rule to itself you end up with some ridiculous answers.
Eg the first 80 percent is completed in 20% of the time. The first 64% (80% of 80%) is completed in 4% (20% of 20%) of the time. The first 51% (80% of 64%) is completed in 0.8% of the total time. This just doesn't ring true for me. It takes less than 1% of the time to complete 50% of a project?
Maybe 60/40 would work better (36%/16%...21%/6%) or 70/30 (49%/9%...34%/3%)?