From: SidTouati@inria.fr   
      
   Well, as in any science, the more you discover, the less you know...   
      
   In compiler optimisation, it depends if you target computer engineering   
   or computer science.   
      
   In computer engineering, you can check the latest or the upcoming   
   architectures of the market (multicore, DSP, VLIW, etc.) and see what   
   you can do for them. Any empirical result would be interesting for   
   quick publication in conferences, but such results become obsolete   
   after few years (even if they would be cited for many years).   
      
   In computer science, you can take an open problem (there are too many,   
   from backend to frontend), often highlighted by computer engineering,   
   and try to have a mathematical result. It is a more difficult process,   
   but your results would stay valid forever.   
      
   S   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)