Programming always gets more difficult. I'm a much better developer than I was 10 years ago and I don't make the same design or development mistakes. But instead of writing perfect code, I'm working on larger and more difficult problems. In fact I'm now responsible for so many more platforms, systems, and programs that I probably spend more time dealing with problems than ever before.
I guess my take on it is different. The things that I've found to be the source of bugs/problems for me are areas that can scale to larger problems. Things like component communication, algorithm edge cases, typos, poor memory management, etc. For me, the error rate slowly goes down, because I've learned to compensate for my weaknesses. Maybe your (and his) weaknesses are not as portable to other and larger problems/systems?
I'm sure my error rate is a fraction what it used to be but I'm producing far more code and working on far larger problems. You better hope that areas of improvement on are the ones that scale to larger problems.
The original article was about error rate, that's what I was talking about:
1. write some code
2. run the code
3. get an error message
4. find the error and back to step 1
Hour by hour, day after day, I do this. Always searching
for what's wrong with what I'm creating, rarely thinking
about what's good about it. It's a negative reinforcement
feedback loop.
If the error rate was lower, steps 3 and 4 would happen less frequently.
My original point was that #1 doesn't stay static -- if it did, then I'd be writing perfect code right now. Instead the code you write gets harder.
Even then, I don't think the author intended you to take that so literally. That's an easy way to describe the problem to someone who isn't a programmer. The negative feedback loop is much longer and more complicated than that.
