Honestly, though, looking more than 10 years back in revision control is usually only for curiosity. Whatever historical reasons existed for code to be a certain way are likely impenetrable or no longer relevant.
Nah, I've often found the opposite to be true. Granted, it requires proper commit messages and code comments, but I've had my fair share of "oooh I see" moments when I encountered nonsensical crazy code. Sometimes it reassured me that it could safely go, other times it revealed a whole new dimension nobody was really aware of anymore that then got a good chunk of comment space in the code base.
A) great commit messages,
B) comments that are worse than the commit message so that you have to go scavenge in revision control,
C) a single or very few commits where the craziness happened,
and D) a codebase that hasn't drifted enough that this signal is hard to find.
I've had poor luck spelunking for reasons in things like the Linux kernel. 90% of my revision control effort looks at things from the past couple of years, and before then it rapidly becomes diminishing returns.
