Thread: Why nothing but beginner questions?

Originally Posted by Dave11

and C++ is nowhere close of dying

Ok. Have been hearing this crap the whole thread, I'll just burst out.

I'm frankly annoyed at how quickly my claim was taken out of context. I expect that sort of stuff on many places, just not here and not by the very people replying. It seems that the mere mention of "C++" and "dying" on the same sentence triggers a defensive mode that just stops you people from reading anything else that was said and make any effort to counter the actual arguments.

"No, C++ is not dying because people use C++". "No C++ is not dying because o many things depends on C++". Right. That's freaking genius!

since C++ is maybe the only language that allows you touch memory and the operating system directly while giving you the possiblity to write Object-Oriented code mixed with smart meta-programing I can't see how C++ can die in near future.
In the end, everything has to reach the OS level, so how will you implement these cool high level features (like coroutines, http etc.) without asking your OS to do it?

yes, there was a huge trend to move from C++ to Java (or C#) back in the early 2000. now let's look what was the arguments back then:
1) C++ is un-safe and leaks memory
2) the code you'll write with Java will be much little than equivilant C++ code.
3) Java has much more utilities than C++

now in 2015 let's review
1) smart pointers, shift of developers thinking no to invent the memory wheel
2) C++14 shreads this claim to dust. actually, many code snippets *today* will be much simpler than Java
3) this also changes as the time goes by. in C++11, we got threads and regular expressions, in further standards we'll get networking, coroutines filesystem library.

so as you can see, aside teh psychological fear, there is nothing really in the claims against C++. C++ is modernizing itself every year withou loosing 1% of functionality or preformance.
I really see no reason why big, highly paid projects wouldn't be written in C++. if it was so - Microsoft wasn't encouraging feveloeprs to write in C++ as side of C# (aka "Going Native" con) and wouldn't develop a whole new framework for windows 10 (aka "WindowsRT").

also, you can't really write triple A games with managed language, it is hard enough to write it with C++, so use a non-deterministic memory deallocation? no way. so in this field C++ is safe for many years to come. also , there are regions where you don't have much options - Operating systems, Networking, Hardwares and many more.

yes, this is not 1995, C++ is not the dominating language anymore. but saying C++ is dying is like saying "English is dying, look! Chinese and Spanish became much more popular!"

Sorry Mario, I wasn't at all trying to be adversarial about it. I just don't agree, that's all...

Originally Posted by Dave11

yes, there was a huge trend to move from C++ to Java (or C#) back in the early 2000. now let's look what was the arguments back then:
1) C++ is un-safe and leaks memory
2) the code you'll write with Java will be much little than equivilant C++ code.
3) Java has much more utilities than C++

now in 2015 let's review
1) smart pointers, shift of developers thinking no to invent the memory wheel
2) C++14 shreads this claim to dust. actually, many code snippets *today* will be much simpler than Java
3) this also changes as the time goes by. in C++11, we got threads and regular expressions, in further standards we'll get networking, coroutines filesystem library.

Completely irrelevant. I don't even know why you speak of Java. The argument is simply that C++ has been giving market share to other languages that are said to be (not necessarily are) safer, or easier to develop on.

There are very few reasons today to code in C++ for the GUI, for instance. Unless C++ remains an important language in your team knowledge pool, for almost all types of projects there is simply no motivation to pursue C++ as a solution. It is much easier to do GUI development in other types of languages. This is why brewbuck argument makes little sense to me. If you are already programming in C++ and you have a strong knowledge base, I trust you won't see a reason to change languages. But conversely, C++ is becoming less and less an attractive solution to these type of projects if you happen to be looking for a language to code your project on.

On the other hand the argument for games really isn't something I care much for. C++, for all its beauty, has always been a big problem in game development. The complexity of language is actually something that has been working actively against development costs and the quality of the final product under such tightly controlled budgets and time constraints, as is the case with the game industry. You can bet many a developer would love to see something else coming up that could replace C++ as a development tool for video game development. I argue that C++ is important in game development, for the lack of a functional alternative. That industry is so full of abstractions, it isn't even fun anymore. All in an attempt to hide away C++ from the developer.

I will argue that C++ or C will always remain important core languages in this architecture. I have always defended the idea (see, the Rust thread) that within the current computer architecture, you just aren't going to find a better alternative for low tier development. But C++ is loosing ground as you move up on the operating system stack. And this is what I am talking about when I say C++ is dying. I tried to clarify that on some post in this thread. Naturally, we can agree to some niche development areas high up in the stack where C or C++ may remain influential. But that's about it.

Programming languages also tend to suffer from a generation bias. The new languages, and the space they have been occupying as they displace C and C++, are breeding new generations of programmers and influencing newcomers. As the old is slowly replaced by the new, you may find in that yet another factor to further constrain C and C++ to a corner of the development ecosystem.

Originally Posted by Nominal Animal

DJBDNS is utterly secure compared to Bind -- bind's security approach would best be described as "oops". Yet, DJB gets basically no respect among programmers, whereas Sendmail and Bind are considered excellent pieces of software!

And the funny thing, a huge security hole in BIND was just recently exposed. A couple of months ago or so.

But for all the overblown egos typical of software development (any engineering field in fact), the truth is that the reason for this inconsistency is that people still operate mostly based on what someone else is saying or doing. You get something like BIND to stick around because a large ego won't ever stop monkeys from imitating other monkeys. You have seen this before, until someone that matters in the industry cares, no one will care. But once that happens, suddenly everyone will fill public forums with how terrible BIND always was.

Originally Posted by Mario F.

As the old is slowly replaced by the new, you may find in that yet another factor to further constrain C and C++ to a corner of the development ecosystem.

I wouldn't look at it as a corner. The finest, most delicate roots are the source of all the water and nutrition for the enormous tree.