About ten years ago I started my first full time programming job. I was working on some business software written in C/C++. I say C/C++ instead of just C++ because it was written mostly in C with classes – with perhaps a few cases of inheritance or actually even only few classes – it used structs in many places to group some methods together the same way you would use classes today. The code was fairly complicated and it was just aging. The standard C libraries it used were not very rich, so if you wanted to code up something fairly straightforward – you had to dig through a lot of documentation looking for something basic only to end up implementing that list class yourself. We were using a Watcom compiler at that time and trying out Borland C++ Builder for some side projects showed it was a great improvement in that VCL provided a lot of great wrapper classes for Windows APIs making things easier. Two years later we were looking into rewriting our application or at least starting a new version and considering which tools to use – basically trying to choose between Borland C++ and C#. I was on the C++ side – it was a fairly known language, more predictable and providing intuitively higher performance, while C# was at version 1.1 and had lots of great libraries, but a bit foreign and why would anyone want to go away from the fun of using pointers, managing object lifetimes and relying on some garbage (collector)? Well, I lost – Microsoft seemed to be focusing on supporting C# and in the end it turned out to be a lot more productive than C++.