[gaius]

Firstly, any real working programmer has days where they write no code. You are writing documentation, meeting with users and sponsors to discuss new features and schedules, doing code-reviews and mentoring, merging branches, debugging race conditions, meeting vendors, chasing dependencies in other parts of the company, a million other things. That brings the average LOC/day down.

I can't imagine anyone sustaining 100 LOC/day over the long term unless they think that cranking out HTML templates or something counts as programming.

[swatow]

I realize that I was referring to number of LOC touched a day, not change to the total codebase size. I'm not sure what the first poster had in mind.

For reference, I count myself as writing about ~40 LOC a day for a greenfield project (which I spend about 50% of my time on), measured by total codebase size.

And no, it is not HTML templates, although it is a fairly verbose language. And I do all the things you describe in your first paragraph.

In my experience, a good programmer is almost always a fast programmer, and non-trivial programs usually require a lot of LOC to get the job done.

[sshumaker]

I have to disagree here. This is definitely possible, especially on more greenfield projects. Most of the engineers on my team produced that much sustained. And my average is closer to 400/day even though a big chunk of my time is spent doing things other than programming (e.g. tech leadership / management). LOC isn't a great metric though and can be easily gamed.