Archive for October 2008
Last update: 2008-10-28.
I usually tend to defend C++ but name lookup is probably one of the most complicated things in C++ to wrap your head around. Namespaces are supposed to make your life easier. They provide means to group certain things together, to avoid name clashes between two parts of your program that might have been developed independently. Since typing fully qualified names is tedious people like to use short cuts: using directives and using declarations. So far so good. How well does this work together with generic programming?
In this post I talk about my programming language transition from Java to C++ and my experiences so far.
The article “fast inverse sqrt” came to my attention. It shows a small function written in C which is amazingly fast and approximates sqrt(1/x) pretty well. Appearently it was used in the Quake source code to speed up vector normalizations. But how does it work? Also, can it be improved?