Thread: why not ++?

>> Converting a project would take almost no effort in most situations, provide future development with much more flexibility while still including the features of straight C

>> and I know everyone who knows C knows C++

Attachment 10197

>> Why do some projects/people do this, what is there to gain?

Apparently your own time by not throwing out an entire code base just because it's not new while it is still well supported on platforms, and there are people alive who can write it; people who think C++ should replace C will stay out of your project and not cause drama.

Originally Posted by User Name:

C is supposedly a little faster, I guess that's a reason...

That is not correct.
C and C++ are tied for speed. C is better in some areas, C++ in others.

Originally Posted by Epy

I don't really think you can say that one is faster than the other when the speed of the code depends on how well the compiler optimizes your code. It's really compiler vs. compiler.

Computer Language Benchmarks Game reflects what you say, i.e. C is faster in only some areas, but take a look at the distributions...somehow C++ has a much, much smaller standard deviation. Since both languages end up as machine code in the end, it's apparent that g++ does a better job of optimizing than gcc.

No, it's not compiler vs compiler. It's about language features.
For example, take the standard library sort. Now, run it, say, 10^9 times. In both C (eg qsort) and C++ (std::sort). Which will be faster? C++, hands down. No doubt about that.

Now, take a hand-written, hand-optimized sort by a guru and compare it to C++'s std::sort. Run it 10^9 times. Which one wins? Probably C in this case.

The thing is that there are some things--like templates--that allow C++ to do things that C cannot. But with C, you usually roll out your own optimized solutions for a specific case, and they are sometimes going to more efficient, giving C the trophy in this case. This could be done in C++, but usually isn't, because we want to achieve more genericism.

Originally Posted by Elysia

But with C, you usually roll out your own optimized solutions for a specific case, and they are sometimes going to more efficient, giving C the trophy in this case. This could be done in C++, but usually isn't, because we want to achieve more genericism.

This can't be done in C since it doesn't share a relationship with C++, the same way C++ does with C. This is an advantage of C++, as indeed I can insert C code in my project and have that specific part be compiled by the C compiler, while the rest is dealt with by the C++ compiler.

However these type of situations are mostly only for academics or for very specialized situations which don't represent the vast majority of programmers' concerns. More commonly we simply don't care which one is faster given that both languages offer a performance window that more than suits the vast majority of tasks.

... and as do many other languages. All the mainstream languages do, for sure.

> and I know everyone who knows C knows C++,
Erm, there's a huge chasm of knowledge between being able to write a C program that will compile with a C++ compiler and actually being able to write a GOOD C++ program with a meaningful class structure.

Nor does the "meaningful class structure" emerge automatically, simply by renaming .c to .cpp.
At best, it's an instant conversion of passable-to-good C into horrible C++. Any C++ native would instantly puke at the sight of it.

As for being trivial, read this -> Incompatibilities Between ISO C and ISO C++

> Remember, I'm just wondering why one would choose to use straight C in the first place
Whoever started the project made a choice, based on what they knew and what they preferred.

With only a little more syntactic hacking, you can turn C into Java. Hell, they both use curly braces and semicolons - it can't be that hard right

Rewriting a good C code is pointless. You should leave the window open for re-writing bad C code in C++ instead of good C code though, if you are going to invest on the revising task.

But there might be still a good reason to re-write good C code in C++ code, just changing it so it compiles with C++ compiler. Then you can continue a project in C++ and use the C code as it is. Using its functions or adding some lines in main. Yes, the code might look strange but still functionality is more important than looks. Of course you have to have a reason to want to use C++. Which might be some libraries for example.

So you have a choice as an example. Continue in C and implement some C++ libraries you want in C or make the C code C++ compatible and continue in C++. The second I would say is a best strategy.

Translating code to another language is a pain (i.e. from Java to C++). Continuing a code to another language might not be. The benefit of C to C++ is that you can use the C code as it with some changes and use C++ for the rest.

C is old, and misses a lot of functionality that is critical to today's requiring applications.
C is a language that is designed for platforms where the language must be easy, fast and cheap to implement.
Today's applications simply require a lot of things that C cannot provide easily. Therefore C++.