I read somewhere that the blockbuster title 'Gears of War' had only about half a million lines of C++ code: ~250,000 specific to the game, and ~250,000 for the entire Unreal 3 engine.<p>I was surprised. Steve Yegge seems to be pretty well known as a programmer, or at least, as a programming blogger, but his 'modest' web game was more than a million lines of Java.<p>Apparently, if you're disciplined enough (they coded in layers, much of which was in a strictly functional style) you can do a lot with a little.
I enjoyed the following at the end of the article (with context added by me):<p><i>Think about that last analogy: A million [lines of code costs a million] times the cost of the flash chips [it is stored on]. Yet accounting screams over each added penny in recurring costs, while chanting the dual mantras "software is free," and "hey, it's only a software change."</i><p>I've worked at a place like that before. They would pay a full team of developers for several months to roll our own custom solution to something we could have bought off the shelf for $1000. People just don't see software as expensive to develop... and unless it's open source, it is.
They need to come up with a new metric. LOC is so wrong for so many reasons. Running tested features seems like a good way to measure project progress but I don't know how it would relate to overall project complexity. <a href="http://www.xprogramming.com/Blog/Page.aspx?display=RunningTestedFeatures" rel="nofollow">http://www.xprogramming.com/Blog/Page.aspx?display=RunningTe...</a>