I wonder - has anyone ever tried the approach of the entire VM binary image being the repo and doing smart bindiffs as changes? The trick would be to identify what is "code/config/program code" vs random/temp files. In code tracking you do it manually by saying don't track. Here you'd need a bit more intelligence, but seems like doable. And you could create then like a git for machines. I may be too simplistic / not…
That's exactly what Cloudlets does. http://github.com/dotcloud/cloudlets
would work?
Wound't the system become a nightmare once I start installing/changing stuff in the machine in a very organic way?
Meaning, I'd have to go and maintain that manifest???