Isn't it about time for D3?

[meppl via Digitalmars-d]

in my opinion starting with D3 is too early, because D-people 
dont even seem to know how exactly to implement new exciting 
features like "reference counting" (and how will "reference 
counting" work together with the new "scope"-feature?).

So, if a D3 gets developed _now_, i guess the compiler developers 
will create syntax issues again for those features and then we 
need a D4. Instead it might be better to collect more experience 
before thinking about D3