Combining Events And Threads For Scalable Network Services This paradigm blending was done in Haskel. Programs that use it have comparable performance to C programs for Disk and TCP/IP throughput and similar performance to Apache WebServer for static page delivery. As well, they were able to launch 10 Million threads in 480 MBytes of memory -- yes that is 48 bytes overhead per thread, they were only able to launch 16 K with the C Native POSIX Thread Library.
How to Use a Kaizen Newspaper I now recommend that people start with learning to see waste and record it. Jon Miller shows how to take this to the next level.
