Restudy C++ and C++11

Oh! C++ … Yes! C++ …

The major reason that I switched to Python from C++ is the complexity of C++ at that time. I thought writing large program using C++ was a disaster: it took too much effort in coding. Instead, if I use Python, I could save tons of lines. However, when I re-read the C++ Primer again today, I know C++ is not as bad as I thought. Especially for C++11.

Programming Styles

Before I start talking about C++11, I would like to tell something about our C++ programming style. Frankly speaking, I am a dis-organized programmer when writing a large project. Usually, I do not take care of header files very much in the small project, but now I think the structure of the whole project is very important for us. Now I am improving myself. For how to organize the your project, I recommend you to read C++ Primer for details. In general, you can put the following (and ONLY) into the headers:

• Function prototypes
• Symbolic constants defined using #define or const
• Structure declarations
• Class declarations
• Template declarations
• Inline functions

C++11

As for C++11, the first feature I like is the tuple type. In Python, it is extremely easy to make a tuple and use it as the return value from a function. However, the previous C++ can drive you crazy if you want to return more than one piece of data, even you can make an object of your return type. Fortunately, C++11 introduces such a wonderful feature. If you like to know more about it, check out this reference.

Then, C++11 surprisingly supports type inference via auto keyword. Just like Python, you do not need to worry about the exact type of the object, the compiler can help you figure it out. But I do not know the exact way in C++11 to derive the type (maybe a semantic approach). The coin still brings a bad edge: obscure programming. If you use auto keyword, it may be hard for others to infer what you really want to write here. I guess that’s why Google C++ Style Standards does NOT recommend us to use it. Personally, it is a good stuff to write your program easy and flexible.

For containers, C++11 adds unordered for set, multiset, map and multimap. In addition, array seems to be very interesting, it provides an interface for C-like array but encapsulated. Comparing to vector in C++, I guess it improves the performance a lot, yet I have not measured it. Further, iterators add the lambda function support. I pretty much like this feature because I hate write for loops by myself. This feature makes C++ much simpler and more concise just like Python’s way.

Finally, the threading. Although it is always a significant component, thread library has not be in C++ standard until this modern one, yet we use pthread for a while. Now you can write multi-threading code as you want, whichever the compiler is. It’s just the standard! Besides threading, the regular expression can also be interesting and important to C++ programmer.

Since C++11 brings so much new features in and achieves an excellent performance which has a large margin comparing to other programming languages (the literature from Google Research). From both performance and simplicity’s viewpoints, I would recommend to use C++11. But I am not saying to replace something, because languages like Python has it’s perfect power to do web service, natural language process, etc. A hybrid language environment is very common nowadays, but we just need to be careful about the right language for the specific work.

P.S.
Check out this Video from Bjarne Stroustrup.