Hello Ion, When Kevlin Henney presented his vision of how the thread interface could be in "More C++ Threading: From Procedural to Generic, by Example" , April 2004 I has very impressed on how the roles were inversed, not only for threads but also for condition variables. When he presented his proposal to the C++ LWG "N1883: Preliminary Threading Library Proposal for TR2" , August 2005 I has very excited. When you announced your implementation based on N1883 [ANN][thread_new] New thread launching library based on N1883 - Ion Gaztañaga May 03, 2006 I started to though that this could be soon a reality. When the Kevlin proposal was not accepted I was really disappointed. Could you point me some links that help me to understand why his proposal was not found interesting or souhaitable? Any way your implementation on top of the thread interface proved that we can have two level interfaces. Independently of the C++ standard, do you think that it will be possible to have two thread interfaces in Boost, one on top of the other? Do you plan to follow your thread implementation, and request a Boost review? I'm really interested, and I supose that I'm not alone. Best regards _____________________ Vicente Juan Botet Escriba