How much Linux kernel code has survived the past five years?

Also see <a href="http://lwn.net/Articles/375550/" rel="nofollow">http://lwn.net/Articles/375550/</a> from the comments and its associated thread.

I wonder if this article will motivate hackers to change/complete some of the particularly old untouched or unfinished code mentioned.<p>Such a technique might be a good way to find older code which could be improved upon or upgraded.

I wonder if most projects would have a similar curve, and if parts of the curve had different properties (bug count, difficulty to change, etc.)