The perception of programming

That's why context switches can be very helpful. Sometimes, when I'm stumped, the best thing I can do is go for a bike ride, chat with my wife, etc. When I return, I might notice something that had been hiding in plain sight.

Brilliant observation.<p>The only difference is that music needs to always stand on its own, so in that case he is always making a mistake. In code, sometimes you are overcooking, but often times you are just reflecting a growing understanding of the big picture. In this case the code may end up quite smelly <i>for very good reason</i>, only to be discovered when the next developer comes along to refactor and only realizes the ingenuity of the original code hours later.

In the first two years I made (short pieces of) music, the things people liked best were those I threw together in an hour or two. It was almost uncannily reverse-proportional to time spent. Probably because it was songwriting, not composing.<p>Programs are like architecture ... each program is like a unique building. But in building, the architect designs and the workmen construct. Too often programs are constructed without design ... many times because you don't really understand the problem until you've worked at it. At which time, like musicians or writers, you have to be willing to throw away the first draft and start all over - this time with a solid architecture.

This is very true but fortunately for us, programming is a little more scientific and objective than making music. When I find myself unsure which route to take I try to fall back on principles such as DRY and coding at the right level of abstraction. At least with programming, you can go back and say "this is a load of rubbish because of X", or "this is and awesome design because of Y".<p>In fact this is what I try to encourage in my fellow team mates. You shouldn't choose a particular design without knowing exactly why it is the better choice over any other.

A very similar thing happens to me when I'm doing visual design.<p>I can work on something for hours and hours and love it. And then I come back an hour later and finally see all of the problems.