[abk]

I was just about to post the same thing. I agree with the underlying thesis that a good shipping product is better than perfect code that never ships, but the author seems to be confusing good code and over-engineering.

Good code does exactly what it needs to, but will be maintainable and won't have to be thrown away when that "fantastic" product ends up getting popular.

I also tend to believe that doing it right doesn't necessarily take much longer. The code I write now is much better than the code I wrote 10 years ago for a number of reasons, but I still get a lot more done, and a lot faster than I did back then.