From: danwang74@gmail.com   
      
   A.L. wrote:   
   {stuff deleted}   
    > Determining what objects could be deallocated  during compilation is   
    > equivalent to "halting problem" that is undecidable. i.e. there is   
    > no algorithm possible that could do this for all possible programs.   
    >   
    > A.L.   
      
      
   Determining what objects can be deallocated at runtime is also   
   undecidable. Reachability is a conservative approximation to liveness.   
   GCed programs can and do still leak, just not as frequently.   
      
   There's quite a bit of memory management you can validate dat compile   
   time. Enough reasoning to build a GC on statically checkable compile   
   time primitives.   
      
   http://www.diku.dk/topps/space2004/space_final/fluet-wang.pdf   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)