Slashdot Mirror
Interview Update With Bjarne Stroustrup On C++0x
An anonymous reader writes "DevX interviewed Bjarne Stroustrup about C++0x, the new C++ standard that is due in 2009. Bjarne Stroustrup has classified the new features into three categories: Concurrency, Libraries and Language. The changes introduced in Concurrency makes C++ more standardized and easy to use on multi-core processors. It is good to see that some of the commonly used libraries are becoming standard (eg: unordered_maps and regex)."
Try writing a large program that needs to do heavy number-crunching in Java/Ruby/Perl/Python, or whatever is your preferred language.
Been there, done that.
Most of the time, the potentially reduced running time of the C++ implementation never comes close to the months saved in development.
And when it does, it's trivial to go in and write the speed-sensitive portions of the program in a faster language.
If you mod me Overrated, you are admitting that you have no penis.
Haha, disregard that, I suck C0x!
hours finding memory leaks? How bad are you at debugging? Plus there are tons of tools to assist with memory leak detection. However, protection is no substitute for abstinence. Learn to write code better. BTW go do some embedded software with C# or Java.
I have C++ code that I maintain. It was written in 2004. I also use Firefox, Notepad++, FileZilla, 7-Zip, which are all written in C++. It seems like most the applications I run are written in C++, with many written in C and some Microsoft programs perhaps written in C#. Java killed C++? You wouldn't be aware of it from the software on my computer.
What a fool believes, he sees, no wise man has the power to reason away.
That's what this is... automatic memory management...bigger libraries... restricting pointers more and more....
I mean, C++ is evolving so badly it makes Pascal suddenly look a lot better as a compile time language.
This is my sig.
Hopefully C++'s revision is better than the crapfest that is "D&D" 4.0.
Because performance is important to some people.
...what do people find so difficult about C++? Use the standard libraries, exception handling, and make sure your news all have deletes, and it's no more difficult than any scripting language. I actually prefer it over scripting languages, which have their place, but feel all sloppy and unspecific. It's like the difference between building a house out of 2x4s and building one out of sticks you found laying on the ground.
I'll consider Java and C# as C++ replacements once they get:
These points are serious, especially the first, without real templates, generic programming/metaprogramming at compile-time is not possible. These two are one of C++'s biggest strenghts, though.
To be fair, C# 3.0 is somewhat nice, especially its functional core. Java is a totally uninteresting language with very small expressiveness. Of course, if the job requires it, there is no discussion, but in my spare time, I prefer C++.
This sig does not contain any SCO code.
I've found that the biggest advantage for C++ is the portability. I have written an application backend for PC's (back in the days of DOS) and since then ported it through various versions of windows, Linux (for web use), Palms, and Pocket PC's.
Using C++ allowed me to very easily make the different processor needs, compatible, by writing little compatibility layers, which would swap bigend values, unpack data structures from disk into memory (so is on even boundary). and so on.
Yes the fast speed was why I originally went with the C/C++ route, but the big benefit has been the portability.
...and roll on the C++-hatred! Second C++ article in a short time, and again lots of venom and anger. "Months saved in development"? Really? What are you doing, implementing your own OS before you start application development? Here's a newsflash: C++ also has support libraries, just like Java, Perl, Python and Ruby. They may not be part of the language specification (and I still think that's a weird idea to begin with, but I'm old-fashioned that way), but that doesn't mean they don't exist.
Anything you could want for in a modern language is there. And nobody is holding a gun to your head and making you write those scary templates if you don't want to.
I'm just positively amazed that Slashdot, in theory home of programmer geeks anywhere, should have such a violent dislike of C++. Not that there is nothing to criticize about it, but it is still an amazingly powerful, versatile tool that programmers anywhere would do well to learn.
You are forgetting about the areas where everything is "unacceptably slow" by default. I.E it's never fast enough. (Real time A/V processing for example)
What is needed in those circumstances is a language that is both "close to the metal" and allows for inline ASM, but also lets you do OO on the structural level.
What are the options there? D maybe?
And I can say with a straight face that you are wrong.
If you base your experiences on pre-2000s C++, you know very little of modern C++. I have been developing in it for more than 10 years, and a few years ago I would have agreed with you, but things have changed. Really.
This sig does not contain any SCO code.
We will see the usual litany of C++ hating here in this thread. The hating will be generally based around misconceptions or problems that are 5 years old.
So to get them out of the way:
If you're leaking memory or spending time managing memory in C++, then you're using C++ wrong. Get a book written in the last 5 years.
If you're worried about compiler compatibility (with the exception of export which isn't much use anyway), get a compiler written in the last 5 years.
If you think that C does some subset of your task better, then write it in the common subset of C and C++ and quit whining. Or, write it in C and link it against your C++ code and quit whining.
If you think that templates simply provide code bloat, then get a compiler newer than 5 years old.
If you think C++ is slower than C, then get a good optimizing compiler (you know one written in the last 5 years) and do a benchmark. You will generally find that templates make C++ faster.
If you think "modern" languages are more expressive, then give "modern" C++ a try (insert comment about recent compilers here).
Sure there are valid complaints about C++, but the majority of them I hear on slashdot are complete bull. The majority of the remaining complaints will be fixed by C++0x.
One remaining problem is the lack of a vast array of standard, business oriented libraries. I don't write business oriented code, and I find the C++ STL one of the best libraries out there since it provides really good support for writing efficient algorithms.
Another problem is the difficulty in parsing C++. Sadly that's never going away.
But if you're going to complain about C++ compared to recent languages here, make sure that you're talking about recent C++ too, and try to make sure the complaints are accurate.
SJW n. One who posts facts.
You do know that you don't have to screw around with any of that in a managed language, right? "Very easily make the different processor needs compatible" my ass--Java/C# do it on their own.
"You can either have software quality or you can have pointer arithmetic, but you cannot have both at the same time."
C++ is an extremely powerful programming language and that is why I use it every day. But it has one major problem: It is too complicated. As long as you do programming full time you are OK but if too much of your time is spent on the application side of things you quickly get in trouble. This is what people like BS don't seem to get - not everyone can spend 100% of their time studying the language.
If it was as good as it stands, then newer languages such as C# wouldn't take off.
Don't get me wrong, I love C++ and it's my primary programming language, but to say it's perfect as it is, is just silly.
+1 IDisagreeSoHeMustBeATrollOrAnAstroturferOrAShill
No, the "premature optimization" thing applies to all areas. Especially areas where it's never fast enough.
Why? It's simple: resource management.
You have X amount of resources to put into your product. X is always finite. It's kind of tough to measure X, but you can think of it as lines of code, man-years, or even just dollars. The amount of resources you have varies a lot depending on your budget, how much time you have, and the quality of the programmers you have. But the important thing is that X is always limited.
Now you have two approaches:
Paradoxically, I hold that #2 will produce a faster program. This is because the X you spend on making the program faster in #2 will be more effective, because you've already laid the groundwork for it. It's always difficult and time consuming to optimize code that doesn't even run yet. It's much more efficient to optimize code that already works. So the result, even though you spend less X on speed, is a faster program.
Think of it as transporting a lot of material into the wilderness somewhere. If you first spend some of your resources on building a road, you'll get the job done for less time and money than if you just start hauling stuff into the woods immediately.
If you mod me Overrated, you are admitting that you have no penis.
You're doing it wrong.
Attention deficit disorder is a complicated issue, spanning several major... HEY LET'S GO RIDE BIKES!
1. Boost.
2. Nonsense. Boost has facilities for this ("any", iirc) and also for something called "sum" types which can achieve what you want in a better way ("variant", iirc).
3. shared_ptr, weak_ptr.
4. Yup. Going to be fixed by C++0x.
5. C++ can be written to be a lot more portable than your Ruby or Python.
6. A matter of taste.
HAND.
And I can say with a straight face that you are wrong.
If you base your experiences on pre-2000s C++, you know very little of modern C++. I have been developing in it for more than 10 years, and a few years ago I would have agreed with you, but things have changed. Really.
Citation needed. What is post-2000's C++? Please enlighten me. All of my professional C++ experience occurred between 1999 and 2006, conforming to the 1998 ISO/IEC spec sitting on my desk, with various modifications made for broken compilers (e.g., VC++6, the lack of support for the export keyword in any C++ compiler I've used, etc.). If there's a later "version" of C++ that is supported by gcc, I have not heard of it.
I did some C++ programming in high school and college, but didn't really dig in until 1999. From that point forward, I spent far too much time building tools---like the aforementioned refcounter, or utilities for atomic access to shared variables across threads---to make up for C++'s shortcomings. After some time, I realized that I was wasting my time and should switch to using a language that comes with these features built-in.
My advice to my junior colleagues is simple: use Python or Ruby (not so much Perl, because of its syntactic ugliness and hacked-up object and exception models) for the control flow, and interface to C when necessary, using open source, peer-reviewed C libraries with existing Python/Ruby interfaces wherever possible. You will not only develop code more quickly, and develop more reliable code with better failure modes; you will make debugging much easier for other engineers who will inevitably have to dig into your code later.
[ home ]
The day when McDonalds can get away with requiring MS CS, MS EE, AND get an abundance of qualified applicants for fry cook positions is rapidly approaching.
Goodbye helloworld.c, hello wantfrieswiththat.cxx
There are vanishingly few programmer geeks left on slashdot. Most of the "programmers" here, these days, are folks who've written a few scripts or set up a movable type install.
There are a few real programmers left here, but they're lost in the noise. You know, the roaring noise made by the python and ruby folks.
This post brought to you by a C++ programmer who happens to love Python and Ruby ( and javascript! it's an amazing language ), but uses the different languages where appropriate.
lorem ipsum, dolor sit amet
I used to think like that, but then all the things you talk about are just syntactic sugar. There is nothing you can do with proper generic that you can't do in Java or C#. Yes, C++ is way more expressive than almost any other language, but that is also its peril.
And when was the last time you used meta programming to solve a concrete problem that could not be elegantly solved otherwise.
Most people learn how to calculate a factorial using meta programming techniques and stop right there. It's more of a curiosity than useful practical technique.
As the island of our knowledge grows, so does the shore of our ignorance.
The more I look at it, Java's Just in time compiler is just about as fast as C++.
The biggest difference tends to be how you program.
When you really learn to program OO, you tend to create many small objects that outlive the object that created them.
In C++, that would mean running thousands of mallocs a second, and having some other random (arbitrary) object delete them later.
That's not the way to code C++. Since you are constantly required to think of object lifetime, you usually have an object die with the method that created it. When you are doing that, you can put it on the stack instead of allocating it on the heap. Stack allocations tend to be much cheaper because there is no deallocation.
But if you actually coded C++ OO, not caring where objects were destroyed, keeping many more small objects around, etc, your performance would be as bad as Java or worse.
For something that doesn't allocate at all--well a simple test was just created on Cedric's weblog that compared a good algorithm in different languages and C++ couldn't match Java's performance until it was compiled with -O3.
For java, apps just keep getting aster (even for programs that have already been deployed). To increase the speed of a deployed C++ program, you need new hardware.
A VM system CAN BE superior to a compiled one in every way since it can be compiled to exactly the same (or much better) assembly code if the VM determines that would be best. Not sure why this whole speed thing keeps coming up. When I see a Java program written to avoid memory allocation, it tends to perform almost exactly like it's C++ counterpart.
I want to like C++, heck, it was the first language I learned. But after so many hours of memory leaks and pointer-induced errors...
perhaps you didn't learn it very well. Check out RAII for one way round your problem, learn about references and destructors for another, and learn about auto_ptr/shared_ptr if you still have difficulties.
If you absolutely must have a GC, put one in. Mono uses this one, so I assume they think its quite good. Stroustrup says that GC has a place in memory management, but it should be the last resort not the first. I agree - if you can't program without getting so lost you lose memory all over the place, GC will not magically help you* as you'll end up losing track of other non-memory resources instead**.
* - unless you're an ex-VB programmer, in which case you need all the help you can get :-)
** - eg. you create an object that opens a file, if you 'leak' that object, the GC will collect it for you when it feels like it, so you may end up trying to reopen that file only to find that its still in use. There's still 'soft leaks' that the GC will not manage where your objects are still referenced even though you think they aren't.
Generally speaking, I agree with your post. However...
That's a really, really awful way to think about C.
C has a completely different set of best practices and design principals; anyone who puts "C/C++" on a resume I'm reviewing loses points as opposed to listing them separately.
"MBA's will not need programmers anymore, so we'll be able to code OSS full time!"
You realize that is what they said when they introduced COBOL, right?
emt 377 emt 4
I do not understand why the post is modded "funny". Indeed, very fast SIMD libraries are coded in assembly. I think even 'liboil' has some parts coded in assembly
I'm just positively amazed that Slashdot, in theory home of programmer geeks anywhere, should have such a violent dislike of C++.
There are at least two ways to interpret that:
Now, "most popular" isn't automatically the same as "best", or else McDonald's make the most delicious hamburgers. Still, if a huge crowd of people who at least hypothetically have experience with the subject matter tend to have the same opinion, perhaps there's something to it.
Not that there is nothing to criticize about it, but it is still an amazingly powerful, versatile tool that programmers anywhere would do well to learn.
It could be that we learned it, cringed in horror, and went back to our more expressive dynamic languages. I have no fear of programming to the metal, but I bought a computer so that I don't have to deal with all the low-level gruntwork on a daily basis. There's nothing more powerful or versatile than raw machine code, but there's a reason why most of us use something else.
Dewey, what part of this looks like authorities should be involved?
It's not that people think that's the only use for them; it's that that's the killer use.
It's a lot easier to persuade people to learn a new feature by saying "this will make your life easier" than by saying "this will let you write better code". Most people don't care how "tight" their code is, as long as it works. What they care about is how easy it is to write and maintain.
I also have an extensive experience with C++, and I tend to agree with a lot of the criticism that it gets.
But the problem is that no alternative exists for the type of problems where C++ is used extensively. I guess the most important area is games.
The world really NEEDS a language (the last low-level language) with the low-level performance of C++/C and with a full, modern library, and modern language features (threading, modern module system (not based on #includes and a crude preprocessor...), optional strong typing system a la Ada with optional runtime-checking etc etc etc.
Basically, a really nice, compiled, well-performing, modern low-level language could easily exist. But it doesn't. So we'll have to settle for C++ until someone makes something better.
An interesting thing is I have written some highly structured code in some rather low level languages. To do so required a skilled architect and an agreement between developers that we were going to honor the rules, even if the complier did not punish us. It can be an expensive way to develop, as competent computer people costs money, which is why it is often more economical to allow the complier to do the thinking.
"She's a scientist and a lesbian. She's not going to let it slide." Orphan Black
If you make language even an idiot can use, idiots will be using it. Like with VB.
Extreme Programming - Redundant Array of Inexpensive Developers
Nothing else in your post either supports, or even directly addresses this assertion.
This is the same bullshit that's trotted out there every time this topic comes up, and it's no more true now than it ever was, which is not at all. If you do a little looking around, you'll find very elegant libraries that support every single feature you'll see in ANY other imperative language, and MANY declarative language features to boot. If you're against code reuse and third party libraries on some sort of general principle, then you're kinda missing the whole point of C++.
I'll be blunt, maybe an ass, maybe a troll, but having used all three of those languages extensively, I can say with almost absolute certainty that the only reason you should be having so much more trouble doing things in C++ (ESPECIALLY as compared to those two languages) is that you're either a very poor C++ programmer or have a pathological aversion to third party libraries.
Another thing that's been very much in the vogue to say lately, but I just haven't seen any meaningful evidence for. I think Bjarne covered this topic fairly even-handedly in TFA, and if he's to be believed then C++ usage is not suffering like popular belief seems to indicate. The crux of it being that web scripting was never a strong domain of C++ in the first place, and in actual applications programming C++ is still the leader of the pack.
std::map::iterator it = m.begin()
That'd be "map<signed short int, unsigned long int>::iterator it = m.begin()". And you can write "using namespace std;" instead of "std::", saving a net minus 15 characters ;)
Well that's just it, C++ is designed to be as general a language as you can manage (which is why so many people don't "get" it or complain that it's complicated because not much is done for you in comparison to other languages) which is why other languages may be better suited to certain tasks.
The example you gave for C# is a pretty good one, but it still highlights that there are areas where C++ isn't well suited to certain tasks and until it is, it's fair to say there's room for improvement.
Thus I welcome any improvements to the language, especially as Bjarne certainly seems to know what he's doing.
+1 IDisagreeSoHeMustBeATrollOrAnAstroturferOrAShill
True today as it was then; i.e. not true at all.
XML causes global warming.
Kay can say anything he likes, but he did not invent the first OO language - it was Simula 67, also a statically typed language, and OOP in C++ has very visible roots in Simula (dot-notation to access members, "virtual" and "protected" originate there, for example; but in general, the very concepts of classes with methods and fields, combined interface/implementation class inheritance, object instances, special syntax for message receiver in a method call, object identity, object references and null, upcasting and downcasting, explicit polymorphism via virtual declarations and overrides, controlled encapsulation by means of visibility levels, etc).
If you make language even an idiot can use, idiots will be using it. Like with VB.
So lets make the language as difficult as possible. That way only good programmers will even be able to write code in it. Never mind the fact that they'll have to spend all their mental effort getting the code to work instead of focusing on the problem they're trying to solve.
The fact that an easy versatile language makes it easy for idiots to program in it is no reason to artificially make a language overly complex. That's insane. It's like making a hammer that requires a PhD to use just to prevent bad handymen from doing handywork.
In other words plenty of good code was written in VB by non-idiots who didn't want to focus on the language but had a practical problem to solve. You can leave the morons to survival of the fittest.
These posts express my own personal views, not those of my employer