Savile Row tailor Thomas Mahon has written a fascinating blog entry on how to draft a pattern. In particular, his story of the need to transcend mechanistic dogma:

This statement utterly terrified me, as we all prefer to have figures and defined points to work with. These had been obtained by a scientific method, so they had to be right, Right?

Wrong. Because what I found out “the expensive way” was that there were times when I had drafted a pattern, checked and double-checked it, and although the measurements were exact, something still looked wrong.

I was blinded by science, not creativity.

This is something everyone in this or any other business has experienced- a gut feeling that you wanted to listen to, but logic wrongly forced you to ignore. Then sadly you’d proceed down this path, and as soon as you saw the results at the suit's first fitting, you knew your gut was right all along, and you have to kick yourself.

In software, it seems we tend to polarize to either shooting from the hip or rigid rationalization. The miracle of software is the fact that software is, like a great suit, soft but with enough structure to give a great fit to the needs of the user.