And those conventions didn't really fit the development model of Pike up until now
could you elaborate on that please? which part of the development model needs a specific layout for branches?
In the devlopment model so far the different pike versions aren't branches, they are new repositories (which just happen to be pre-populated from old ones).