An interesting discussion of the issues involved in improving a
language, pretty relevant to D.
That blog post doesn't list the costs for Python of that stasis, some developer
may leave looking for more interesting grassy fields to work on.
I agree that D2 has to give time to people to create a D2 compiler with llvm
back-end, and I agree that D2 needs some time to refine its implementation: to
remove bugs, write documentation, clean semantics, remove sharp corners,
battle-test the standard library, finish to implement correctly some of the
planned but not fully implemented features, improve efficiency and speed of
several details, refractor the front-end into two layers to allow it better
grafting on the llvm backend, and so on.
But D is not as widespread as Python and it's not as polished & smooth as
Python 2.6, so some small changes may turn out to be necessary, for example to
remove some of the design mistakes.