Thread: Why C Matters

I know a will be equal to be then

No1 - In C you know that this 2 objects will be equal
in C++ you need to read the code of the Copy constructor or operator = to know what is going on

And just looking on the code - you do not know if you need to read the additional code or not.

In C - when you see
MycopyStruct(&a,&b);

you know exactly what function is called and why.

Originally Posted by abachler

The assumption is that it won't be an everyday thing. The transition period is usually about 1 month before you leave, and up to 2 months after you leave, and the severance package should reflect this. No severance package, then of course no support, but then its not what I would call 'amicable'.

That what I said - when found a bug in a code that was written a year, or two, or 5 years ago - you on your own. You will take a code and have to figure out by yourself what the person that wrote the code has thinking while writing it.

Originally Posted by vart

No1 - In C you know that this 2 objects will be equal

But is it correct? If you need a deep copy?

in C++ you need to read the code of the Copy constructor or operator = to know what is going on

Sure, but if the copy constructor does something else than the necessary work for copying the object, then the copy constructor is flawed and you have a bug in your code. I should note that I should be able to know that when I see a = b, regardless of how the class is implemented, a SHOULD be equal to b afterwards.

And just looking on the code - you do not know if you need to read the additional code or not.

In C - when you see
MycopyStruct(&a,&b);

you know exactly what function is called and why.

In that case, the C++ code is no more difficult than the C code.

But how are you going around polymorphism in C? I don't know, but I do believe it will look messier in C than it would C++.
Perhaps you might argue it's harder to understand WTF is going on, but the code is cleaner, nicer and you just need to understand what the objects does.
This is what I believe anyway.

But is it correct? If you need a deep copy?

Logical correctness is independent of the programming language. Whether you need a deep copy or not has nothing to do with how a deep copy is done.
The point is that in C, if a deep copy is done, it is done very explicitly. In C++, you have to look up the constructor - either the code or the documentation - to find out whether it does deep or shallow copying.

Actually, since most objects have a "sensible" copying mode, that makes it actually easier for casual reading. But I'm talking about auditing, about tracing exactly what the code does and whether it conforms to specific criteria (security, execution time, ...).

Yes, well, then again... if all objects handle themselves, then you don't really need to know their specifics - just use their implemented interface! That is typically how classes should work. So that can take away some time.
But if you need to trace every single line, sure, it will probably take longer time...

Originally Posted by vart

That what I said - when found a bug in a code that was written a year, or two, or 5 years ago - you on your own. You will take a code and have to figure out by yourself what the person that wrote the code has thinking while writing it.

Personally I would refuse unless it was a fairly straight forward piece of code. Ive done this before. I usually just tell my boss, 'hey, its spaghetti code, it will be faster to write a new function from scratch'. I only guarantee tht my code will compile and run, eventually. i dont guarantee that it can be maintained by anyone but myself. In fact, I dont even guarantee that. My boss gave me a new feature to add to a program I havent worked on in 2 months, adn its to a part of the program I havent messed with in 6 months. Its going to take me a day and a half just to refamiliarize myself with the program, and I WROTE IT. The cold hard fact is that the more complex the project, the less likely the code will be eternally maintainable. This is doubly true if the initial project isnt managed properly.

i dont guarantee that it can be maintained by anyone but myself. In fact, I dont even guarantee that.

And that's why I wouldn't hire you.

Originally Posted by CornedBee

And that's why I wouldn't hire you.

I doubt very much you work on projects that would require my skills or that would even interest me, no offense. Most of my code is so highly optimized that even I find it difficult to read after a few months. A subroutine that can process a 30 fps MJPEG stream on less than 2% cpu tends to not be very readable.

I very much doubt I'm in a position to even hire anyone.

But still, your attitude is something I cannot understand.

And documentation. Which is why we use them -- right?

@abachler: So are you saying that we shouldn't even bother trying to write good code because all software projects are doomed in the end? Seems a bit fatalistic and pessimistic to me.