On 17/12/2013 03:57, Quoth Andrey Semashev:
I think the transition to git, which is internal to Boost, is not enough to warrant the major version number increment. Such version change typically means major changes in the public interface or other properties visible by users.
Transition to git by itself does not seem sufficient to warrant v2.0, at least to me. Modularisation *might* be, depending on the consequences of this to end-users (those who only use the tarballs). I don't know what the plan is there, but if Boost is only going to be released as a single monolithic tarball as before then I don't think this warrants v2.0 either. If it's actually going to be released modularly as well (eg. tarball per library, letting the user select the subset to install) then that might be enough, in my opinion. Of course the C++11-only version ought to be another major version bump, whether that's the "real" 2.0 or pushed out to 3.0.