Author: Landon Rabern
Date: 16:18:02 08/18/00
Go up one level in this thread
On August 16, 2000 at 20:23:50, Vincent Diepeveen wrote:
>On August 16, 2000 at 19:41:46, Dann Corbit wrote:
>
>>On August 16, 2000 at 18:56:15, Tom Kerrigan wrote:
>>
>>>C++ is a superset of C.
>>
>>This is largely true, but there are enough exceptions to make a remark.
>>
>>There are a very large number of legal C programs which cannot be compiled as
>>C++. There are C and C++ programs which will operate differently, depending
>>upon whether you are compiling as C or C++. Trivial examples:
>
>Those examples are not real true. They form the difference between
>an OOP language and an imperative language. Therefore the difference is
>logical. The core of a chess program however compiles perfectly fine.
>
>In fact what i did to the auto232 player as patched by Remi Coulon,
>i renamed a few 'cout' commands to printt, renamed the c++ file to c and
>it worked at once.
>
>If you have an interface that's doing all kind of things like putting
>multiple boards on the screen, then if you write a very neat way of C,
>like nice modules, it's no problem to do this in C. If you write in C++
>then with less experience it's easy to make objects that show
>multiple boards.
>
>Personally the only 2 big advantages of c++ which i miss a bit in C
>are
> - call by reference is real cool
> - data-hiding if programs get bigger that really is cool,
> especially the first few years i programmed in C i suffered from
> this.
>
>What i dislike of C++
> - you need to write hell of a lot more code to do the same,
> now hopefully that's because i'm bad in C++.
> - a simplistic c++ program where you use a few of the c++ features
> compiles to incredible sizes, and it's dead slow. For example i
> planned to write the datastructure in the diep-win32 interface
> in c++, so that the gamedatabase could use a nice OOP methods,
> but after a bit of testing it appeared that using nice c++ it would
> get incredible slow.
>
> Now i program it in C, and after some toying with a lot of typedefs
> and structures i really like that way more now.
>
>Personally i don't see however why someone who starts to program now,
>would NOT start with C++. As you learn
> a) c++ very good
> b) you learn tricks to prevent the above slowness of c++ to happen if
> you need it.
>
>So personally i would advice to start with C++ directly. C is for dinosaurs
>from the previous century, who were born before or short after 1970.
I was born in 1981 and I started with and still use C. I know C++, but I think
OOP is not necessary for a chess program, especially one with a single author.
-Landon
>
>>/* Perfectly legal C, but not legal C++ */
>>/* Hint: just changing the variable names to non C++ keywords is not enough to
>>fix this program to make it legal C++ */
>>#include <stdlib.h>
>>int main(void)
>>{
>> int *new = malloc(5);
>> int delete = 0;
>> return delete;
>>}
>>
>>/* Legal in C and C++, but answers are different: */
>>#include <stdio.h>
>>int main(void)
>>{
>>printf("sizeof the letter 'a' is %d\n", sizeof('a'));
>>return 0;
>>}
>>
>>>You will be able to write your chess program in C, compile it with a C++
>>>compiler, and get exactly the same performance.
>>>
>>>If you use C++ features, your performance will decrease depending on which
>>>features you use and how often you use them. The performance hit can range from
>>>trivial to extreme.
>>>
>>>In my opinion, the goal of OOP is to organize and simplify complicated things. I
>>>don't think a chess program is so complicated that it can benefit from OOP.
>>
>>I do agree with your general premise.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.