> >Yes, ex-Be engineer, and probably ex-Be engineer forever. > Ouch!, that line sounded bitter. I hope not forever ! [Uncle Sam pointing his finger /] [Uncle Sam voice] WE NEED YOU ! [/Uncle Sam voice] :)) > I think the article shows that people have different tastes. > ie: Multithreading vs. Non. Well I'm more used to monotask programming in C (even under UNIX), but I begun C++ under BeOS 2 months ago and I really liked it: http://clapcrest.free.fr/revol/shot_nplay2.png (someone said multithreading ? Hugh :)) ) I also have some experience on multiprocessing on the kernel side: http://lpg.ticalc.org/prj_prosit/index.html I mean processes appear as threads for the kernel anyway, and this is even more obvious on platforms like this one which lacks an MMU. :-D I don't know if you can imagine the feeling you have when you see all those user proggies running smoothly altogether, and knowing that "Yeah they run thanks to what I coded" :)) But I really understand some people feel confused by all those concepts and obscure terms (I'm still not sure I know what a spinlock is, or maybe I already know but I never put this name on it). > Personally I'm quite comfortable debugging multithreaded apps. Debugging also changes some habits. > The concept of parallelism will become very important as operating > systems and programs become more distributed. This is already > happening in business and science applications. > A multithreaded OS starts us off on the right foot as far as I can > see. > This kind of multithreading also lends itself to the even distribution > of workload across multiple processors, which BeOS excels at. When will we see BeOS based clusters btw ? It would rock ! François.