When I was reading A Conversation with Jeff Bonwick and Bill Moore in the latest issue of the ACM Queue there was a passage that really touched my heart of software engineer:

PAWEL JAKUB DAWIDEK At first I just wanted to see how much work it would take to port ZFS to FreeBSD. I started by making it compile on FreeBSD, and once I did that, I was quite sure it would take at least six months to have the first prototype working. The funny thing was that after another week or so, ZFS was running on my test machine. It was truly surprising that the code was so portable; it was self-contained and I had initial read-write support after 10 days of work.

Isn't it really the essence of internal quality of a software system?

You may disagree but to me software product|design|implementation|etc. is great when you get something that you originally did not plan to get. This happens when you do not compromise a single bit of quality of what you do even it is not stated in the requirements. Quality always pays-off, but you have to invest in developing "101% quality" mindset in you team before you get dividends.