I'm totally going for Baghdad :cool: EDIT: Oh man, it's not even there! :sigh:

I agree :omg:

I agree on the templates. In C++ the type system is more or less Turing complete - in C# you just get generics. Also, C++ has multiple inheritance, C# does not. I know this was a design decision but more than once I have started something in C# and switched back to C++ for lack of this mechanism. The perception that C# takes less time to write programs in generally just comes down to the library you get. If a really good, object-oriented framework were released for C++ I think some people would realize just how simple it can be. The difficult thing is creating the tools that can be used to get there. In terms of GC, in my experience manual deletion doesn't get too out-of-hand. In terms of noticeably slowing down development I'd say more time would be spent deciding what path to take to solve a tricky problem. Most memory management can be integrated into classes that know what kind of data they're dealing with. The real problem here is that too many libraries don't leverage this utility - personally I hate calling something like RegQueryValueEx with null parameters, allocating a string then calling it again just to map a registry path to a value. Surely the system should be able to do this itself, after all there are plenty of windows API functions for memory management. In talking about power, some things do expand the power of C++. The yield stuff, IMO, actually does give C# something C++ doesn't have. Assuming this is implemented how I would expect (context switches/coroutines/fibers), this allows efficiency to be achieved in some situations, in that state information for your collection or whatever is maintained implicitly on another stack. Abstracting this in C++ is possible, but needs quite a bit of effort, especially with the Win32 CreateFiber, etc. routines. As for anonymous functions and LINQ, I don't think they really make C# more powerful. LINQ certainly not. I think it's a horrible idea - integrating an application-specific feature into what is supposed to be a standardized programming language, which in essence just redirects your queries and such - does not make it more powerful at all. Anonymous functions add a little bit I think, and are certainly useful - but what they usually do can be achieved in other ways, with just a bit more boilerplate code. Lambda functors have been implemented in C++ with some utility (in Boost, unsurprisingly), so this becomes trivialized. That C++ is able to implement things like this in a library (albeit somewhat awkwardly), and C# (to date) can't or won't show