This post reflects an observation extending the "Rational" and "Use Cases" for employing co-located branches mentioned here: http://doc.bazaar.canonical.com/developers/colocated-branches.html . Essentially, branch co-location permits a new project to borrow artifacts (code, documentation...) from existing and evolving ones to form the basis of a new project. These branches from one or more existing projects and a branch dedicated to the artifacts specific to the new project's functionality comingle to form a unified working directory representing the new project. For example, a group could borrow the "best" code from various remote desktop VNC implementations and create a best in breed derivative VNC. I liken this process to genetic engineering where you take slices of DNA, from various hosts that represent the traits you wish to express then recombine them to hopefully form a viable life form that reflects the selected comingled traits.