Knowing C++ Beyond a Beginner Level

Nerval's Lobster writes: C++ is not an easy language to master, but many people are able to work in it just fine without being a 'guru' or anything along those lines. That being said, what separates C++ beginners from those with 'intermediate' skills, or even masters? According to this Dice article, it comes down to knowledge of several things, including copy constructors, virtual functions, how to handle memory leaks, the intricacies of casting, Lambda functions for C++11, (safe) exception handling and much more. All that being said, is there one particular thing or point that separates learners from masters?

Comments

Knowing when not to

[perpenso]

What separates C++ beginners from those with 'intermediate' skills, or even masters?

Knowing when not to use templates, virtualization, [insert favorite c++ function here], etc.

Basically knowing enough about programming and problem solving with a particular language to tell a need from a want. Needing to use some language feature vs wanting to use some language feature. And being mature enough to stick to needs rather than indulge wants.

Or to state things differently ... all the features have a time and place, and its probably not all the time and in every place.

Re:Knowing when not to

[PolygamousRanchKid]

I agree with your comment entirely. I would only like to add that a true C++ Master writes code that a C++ Novice can understand.

Time to get philosophical. Tomorrow, you could get run over by a bus. Take a wander around your Cubical Town. Are there enough folks there who could take over ownership of your code?

You can do some really cool things with C++. But if other folks cannot understand them, well it's best not to do it. Cool C++ features are like nuclear weapons: very powerful, but think about the consequences of using them . . .

Re:Knowing when not to

[eulernet]

You are thinking like a manager.
As a programmer, I don't want to be replaced easily, and I don't care about my work when I'll die, or even when I quit my company.
I have no problem to share my knowledge with my co-workers, but why should I write code for somebody who'll replace me ?

Also, unless you write frameworks, I doubt very much that your code will be reused.
It will probably be rewritten.

Re: Knowing when not to

[Anonymous Coward]

but why should I write code for somebody who'll replace me ?

Because future you will replace present you. If you need to revisit code years later to fix a bug or add a feature, you want to be able to pick it up straight away, rather than try to figure out the mess you left behind.

Re:Knowing when not to

[TheRaven64]

making the code run faster qualifies if it's not running fast enough yet

And then only if backed up by realistic macrobenchmarks. There are a lot of things in C++ that have interesting performance characteristics (templates allow more inlining, so make microbenchmarks faster but can cause you to run out of i-cache and make the whole program slower, virtual functions prevent inlining unless the compiler can do devirtualisation, but are actually slightly faster to call than cross-library calls via a PLT if they're not inlined). Generally, algorithmic improvements will make a bigger difference than any language feature. The main reason for using templates should be to eliminate copy-and-paste coding, not to improve performance.

Re:Knowing when not to

[TheRaven64]

If you can't be replaced, then you can't be promoted. Do you really want to be maintaining the same program for the rest of your life? And do you want to have a reference that says 'no one can understand this guy's code' when you leave for the next job?

Re:Knowing when not to

[serviscope_minor]

This is complete and utter tosh.

No one I know who does high performance code (such as numerics, real time computer vision, that sort of thing) uses anything but C++, especially for new projects. There is nothing out there that combines the speed and expressivity of C++, and when you know performance is going to be a factor at some point, C++ is the only choice.

Frankly for a lot of scientific and numerics style coding, I often reach for C++ initially even when performance isn't required since it often maps on to those problem domains better than any other languages I've used.

Oh and then there's the embedded world, where your choices are C and C++. Plenty of people use C++, since it like C scales all the way down to the 8 bitters like Atmel.

Re:Knowing when not to

[serviscope_minor]

If you need really high performance you don't use most of the C++ features anyway, and end up basically writing straight C.

Nope nope nope nope nope nope nope nope nope.

That is far, far, far away from correct.

Check out something like Eigen or TooN (a somewhat more obscure library which is used in the vision world for things like PTAM). They are very far from C++. The code written down in an editor reads more like maths. There's no explict loops, no explicit memory allocation. They're both high performance libraries used in challenging applications (seriously download and run PTAM, it's amazing).

They're also fully templated so you can stick in an AD type instead of a normal number and get derivatives out automatically with no extra effort.

Writing high performance C++ is nothing like writing high performance C. It's much better. All the messy C details of allocation and etc simply vanish, leaving clean, nice looking code which is straightforward to read, write and debug.

Or another example from today for me. I need to find the best N (lowest cost) elements during some wort of complex search. It's easy. Just create a priority_queue<pair<double, int>> where the double is the cost and the int is the index of the object.

Then about 3 lines of logic keeps the pqueue updated with the best N so far.

All the irrelevant logic of how a queue works is witten and debugged by someone else and hidden from me without ever losing either performance or type safety. For bonus points, if I find I'm searching very small things and memory allocation becomes a penalty, I can switch the entire thing to stack allocation with almost no effort at all!

Re:Knowing when not to

[eulernet]

If you can't be replaced, then you can't be promoted.

It depends on the kind of promotion.
Not everybody dreams to become a project manager.

Do you really want to be maintaining the same program for the rest of your life?

Not really, but if the pay is good and the job is nice, why not ?
Personally, I have a life outside of my work, so I don't really mind.

And do you want to have a reference that says 'no one can understand this guy's code' when you leave for the next job?

That's the least of my worries !
Do you think that the guy who will take your place won't hate you, even if your code is beautiful ?
Do you believe that your company will not fire you if there are problems ?
If the company doesn't care about me, why should I be faithful ?

Finally, I have a personal question: why do you work ?
Is it to receive aknowledgment, money, fame, self-esteem, or something else ?

Re: Knowing when not to

[eulernet]

I have 30 years of coding experience.
Even though I believe I'm quite a good coder, when I read code from 5 years ago, I'm always surprised to realize that I can do better and simpler.

Whatever the state of your code is today, it will be a mess in a few years.

Re: Knowing when not to

[evenmoreconfused]

Even though I believe I'm quite a good coder, when I read code from 5 years ago, I'm always surprised to realize that I can do better and simpler.

Whatever the state of your code is today, it will be a mess in a few years.

Yes. I find this to be the case too (I was a programmer in the seventies and eighties, and have been in programming management ever since).

So it begs the question "can one write code that one won't think is sub-optimal five years from now?". I've begun to suspect that one can't -- so just learn to accept it and move on.

What I try to get programmers to do is write code that is A) clear and simple and B) balanced in terms of development vs. maintenance time. I don't want programmers wasting time perfecting code that's not going even be looked at for years to come, nor do I want code that takes days to get into when attempting small fixes.

It's like building a house: if you follow the building standards, it's quick, safe, and any plumber or other trade can walk in later and quickly fix or modify things. If you do a bodge job it all has to be torn out and redone properly, or, if you create custom installations, it gets very expensive to create and especially to maintain.

Re:Knowing when not to

[Actually, I do RTFA]

You are thinking like a manager...I don't care about my work when I'll die, or even when I quit my company.

I'm both a programmer and a manager, so I can probably weigh in. I do care about what happens to your work when you leave the company. And part of my job is to make sure that your code is usable.

why should I write code for somebody who'll replace me ?

Because code is an asset that you are being paid to create. And if your code is not maintainable, it's not much of an asset; it's worth far less to me. So, if I notice our code is a fuck storm of uncommented overly complex verbiage, I'm not letting you work on anything important. So, maybe you get to work on small one-offs (really career enhancing), maybe I just fire you. But certainly I don't expect to allow you to keep extending your tentacles..

The reason you say "most of your code will be rewritten" is because it sounds like your code is poor. Why would I pay asset-level prices for stuff with a shelf-life of a year?

Re: Knowing when not to

[crgrace]

Wow. You're right. It's only been four hours and your comment is already a mess.

Error Handling

[oggiejnr]

I've always considered error handling to be the most important thing when it comes to knowing a language beyond the beginner level. Every language has it's own idiomatic ways from RAII in C++, finally/using in Java to the myriad of ways of handling return codes in C. It is also frequently undertaught in most programming language courses.

It is for this reason I despise seeing C/C++ on CVs. It implies that you don't have a strong foundation in either language as idiomatic code is so different between the two. By all means list them as two separate languages, but be willing to demonstrate sound knowledge of both, not the bastardised, resource leaking hybrid I see so often when the term C/C++ is invoked.

Re:Error Handling

[Ambassador Kosh]

I may hate seeing it written that way also but that is also the current most common way to write it and that is what HR systems and their computer screening systems except. You are writing a resume for an HR person usually and if you want it t be seen by someone more important it has to get past them first. That means doing stuff that HR people understand.

There are many fights to take on in this world and others that are just not worth the effort to fight since the costs are so high compared to the benefits and this is one of those fights.

It's not about knowing, it's about understanding

[rippeltippel]

Since C++ is the language of choice when you need performance (along with C and - sometimes - assembly), to write good code it's essential to understand what each line of code does to the machine (memory, registers, ...) and if/how instructions can be optimized by the compiler.

This level of awareness is generally not required to be proficient in other languages, but in my experience it's what makes the difference between newbies and pros, at least in the areas where C++ is used for a good reason.

Said that, it can be useful to understand as much as possible of any language and C++ can provide strong foundations in that sense, as this short article points out: http://www.forbes.com/sites/qu....

Warnings

[wienerschnizzel]

There is no single one thing to look for, because what you really want is EXPERIENCE with C++ programming and that encompasses a lot of things. But if you want an easy test to see if somebody has got enough experience, get some code sample that produces a lot of different warnings at compilation time and have them explain what the warnings mean and how one should get rid of them.

I would suggest the stl

[Ambassador Kosh]

There are many things you can use to improve your c++ code like std::vector. With that you store data contiguously in memory but you also don't have any manual memory management. No new, no delete, no malloc, no free.

For my high performance work I tend to use std::vector, BLAS and LAPACK and my programs usually have no manual memory management of any kind in them. Valgrind shows no memory leaks and the programs are very easy to read and work with.

If you want to do high performance c++ then learn OpenMP and MPI. If you want to do threading just use OpenMP since that makes it VASTLY easier to get threading correct. Add tasks with OpenMP along with their dependencies and you end up with a nice cross platform and very high performance code base if you have done your job correctly. If you need to scale to multiple nodes then use MPI between nodes.

Re:Masters know their limitations.

[jonwil]

I thought I was a guru level C++ programmer (I have been programming C++ for around 2 decades) but even I am constantly finding out about new things I didn't know existed.

Non-master

[gnasher719]

I recommend this link: http://programmers.stackexchan...

There is someone asking whether 8 lines of slightly clumsy looking code can be replaced with something better. The beginner wouldn't ask that question and wouldn't know an answer. The master would say "your code is just fine", because it is actually straightforward, easy to understand, easy to check for correctness. The first answer on stackexchange adds two arrays, one 20 line function, and a few lines of function calls resulting in code that is hard to understand and verify.

Now where C++ is a bit unfortunate is the fact that once you leave beginner level and think you know it all, you have unlimited potential to create code that nobody can understand.

Re:It's not about knowing, it's about understandin

[rippeltippel]

so, you claim you know:
...
int a = 13;

In which register a is residing, supposing we are on an ARM? Or suppose we are in an 68k? Or suppose we are on an x86?

You are just an idiot!

I don't claim i know, I claim that I understand: something that you're clearly neither willing nor capable of doing, as your question and offensive language suggest.

Have a nice day!

Masters don't play with Dice

[MightyDrunken]

One of the defining feature of a non-beginner programmers is that they don't read Dice articles to find out anything about computer languages.

Re:Experts...

[Rei]

Memory objects that clean themselves up after they go out of scope are the devil's work. The devil, I say!

Who else stopped reading at "this dice article"?

[Qbertino]

Just asking. :-)

Re:Given how C++ is taught.

[Rei]

"Smart pointers" are great -- if you don't care about performance (in which case, why are you not using Java?).

Since when does Java's performance even come close to C++'s in benchmarks? C++ performance is generally very close to that of C's, and in some cases exceeds it (example: qsort vs. std::sort - C++'s use of templating allows for inlining of the sort function code)

Smart pointers have very, very little overhead. The worst is std::shared_ptr, and it's still only adding a reference counter, and that's only used on pointer copy and deletion. And if you have a use case that requires std::shared pointer as your smart pointer of choice, then this is counting that you'd have to be doing anyway in some form or another.

  From the benchmarks I've seen, most people see about an additional 5%-ish overhead in debug mode with std::shared_ptr vs. raw pointers in pointer-heavy code. In a release build there's generally no measurable effect (the difference being, in debug mode it can't inline the dereferences).

Re:Simple ...

[serviscope_minor]

Wow the smug condesention is strong in this one.

I for my part wrote an STL clone around 1993 when the STL was just a lab experiment at HP.

The hard bit about the STL is the whole concept of, well, concepts that Stepanov finally hammered out. The STL, especially 1993 era is not all that complex.

Well, iostreams and their interaction with locales is deeply fiddly and I'd steer clear of that. But the basic algorithms, you know, vector, list, set/ma/multiset/multimap, sort, heap, priority_queue and so on and so forth are not too bad.

Not to say it's not an achievement, but it's not enough to convince me that you're an uber-guru. I've written STL compatible containers, and STL like algorithms for things that weren't in there. Apart from quantity the principle is the same.

Perhaps you should read what I wrote: I have roughly 15 years consecutive C++ experience from 1989 till 2005, plus random 3 or 4 years over the last decade.

15 years experience, or 1 year of experience repeated 15 different times?

Given you've never seen code without new in it (as your other post claimed), I'm inclined to think the latter because you seem to be deeply ignorant of whole swathes of C++ style. In a lot of code, you never see new and delete. Everything is managed by containers. I work on computer vision systems, and you can get entire working, robust systems without a new anywhere in sight. The custom containers might have a new in, but that's---well, let me check the library I'm using---let's see there's 80 instances of new in 40k lines. Of those most are in old code from before TR1 gave us many standardised smart pointers, and others could easily be replaced with a std::vector (the code's not perfect, it's been hacked on for the last 15 years), some are strange, silly uses and the rest are to initialise now deprecated auto_ptr.

There's about one legit one which uses placement new and posix_memalign.

With spare time, I could make that one in 40k lines of code easily. In fact that's going to happen slowly as the library is being transitioned to C++14.

I find it terrifying that someone who pust themselves forward as a super C++ guru is splattering new so much all over the place. But you won't believe me because without knowing anything about me you've convinced yourself that you're superior.

Let's see what a real, certifiable C++ guru says:

http://www.informit.com/articl...

Bjarne doesn't like new/delete either. No offence, but I'll take his invention ans stewardship of the language over your 1 year of experience repeated 15 times.

I doubt you regularly find one here on /. who has significantly more C++ experience.

Out of interest, do you have any T-shirts with disparaging things written about n00bs on them? And are those slogans visible under the cheeto dust?

Re:which part

[MiskatonicAcademic]

Scientific (matlab but faster): who cares, you just want the answer, not the software, right?

Not always true. Sure, there is plenty of well-motivated ad-hoc coding in scientific research. However, we sometimes have supercomputers working for months to generate the answer. Even with well-written software this could mean many core-years of number crunching. Without good high-performing software we would not get the answer at all. Developing good scientific software takes time and effort too, but if the software can be used over and over to efficiently solve >1000 problems (for instance, the GROMACS papers have been cited by users ~15,000 times), then the time invested can be very good use of taxpayer money. C++ is not a bad choice for such software in that it enables very good performance and decent maintainability.

Re:Given how C++ is taught.

[TheRaven64]

"Smart pointers" are great -- if you don't care about performance (in which case, why are you not using Java?). If the object is owned by one thread, it's just sloppy to put in the overhead of smart pointers to make your life easy.

If an object has a single unique owner, then std::unique_ptr has no run-time overhead and will ensure that the object is correctly deleted even in the presence of exceptions. I agree with the grandparent: any object that isn't allocated on the stack should be created with std::make_shared or std::make_unique.

Re:Masters know their limitations.

[UnknownSoldier]

This 100%. C++ has become a clusterfuck of over-engineering and I say that as someone who has worked on a C++ compiler.

* /Oblg. Comedy: Hitler on C++

When you have even committee members admit they only use a sub-set then you know the language is too big.

The C++ committee recognizes there are many problems with C++ iostreams but nothing is being done towards performance and type safety.

The committee would rather argue over the rare case of multi-dispatch / multi-methods then fix core issues.

* http://www.stroustrup.com/mult...

Crap like long long, "long double", etc. should have been deprecated in year X, and removed in year X+5. Are they going to add "long long long" someday?? Having types like "double" in 2015 is just retarded -- replace it with "float64_t", and the fore mentioned long double with the clear "float80_t". Bandaging the problem like int_fast32_t doesn't solve anything. How many fucking integers types does the compiler need to throw at us?? short, long, long long, int, long int, int_fast32_t, int_least32_t, etc. and I'm not even talking about MS's hacks of __int32, __int64, etc. Simplify the dam language already!!! Set year 2020 as the date when these barbaric types are deprecated, and year 2030 when they are removed.

Modules have been in a constant state of on-again-off-again for over 10 years:

* First mention N2073 (Sept.2006)
* Revived N4047 (May 2014)
* 2nd draft N4214 (Oct. 2014)
* 3rd draft N4465 (April 2015)
* Wording N4466 (April 2015)

The pre-processor is STILL broken. One would expect #define token operation to work for ALL user-defined tokens. i.e. This isn't rocket science, just a basic Search-and-Replace:

#define @(func) printf("%s\n",func)
 
void foo()
{
  @(__func__);
}
 
// Hell, even this should work
#define @ printf("LOL. Your pre-processor is broken. HA-HA!\n" )

There are no standard pre-processor macros for function names as a string. GCC has the excellent __func__ which Microsoft finally got around to implementing C99 N2340 in Visual Studio 2015!

The C++ committee failed to learn the first lesson about design:

* "Needlessly complexity is a symptom of bad design."

Or paraphrased from Einstein:

* "Things should be as simple as possible, but no simpler"

Re:Masters know their limitations.

[serviscope_minor]

This 100%. C++ has become a clusterfuck of over-engineering

Nope. It's mostly hampered by the need to be backwards compatible. The committee know full well if they make serious breaking changes, then the C++ world will split into two distinct languages. Many, many people won't adopt the new one and the vendors will continue to support them because there's money there.

It's a blessing and a curse. The curse it it's hard to do anything but add features. The blessing is my code I debugged 13 years ago still works.

The committee would rather argue over the rare case of multi-dispatch / multi-methods then fix core issues.

You're basically being an arse. You're picking out one rather speculative paper designed to help C++ stay modern a ways in the future (and somehow labelling this as a bad thing) and yet knowingly ignoring the many papers on fixing core issues.

Modules have been in a constant state of on-again-off-again for over 10 years:

Yeah they should just slap in some module crap and if it's wrong, just rip it out and replace it. Who cares about getting things right and not breaking working code, eh? They learned their lesson very hard with export templates and are keen to not have to learn the same lesson again.

There are no standard pre-processor macros for function names as a string. GCC has the excellent __func__ which Microsoft finally got around to implementing C99 N2340 in Visual Studio 2015!

So you submitted a proposal, right? Or do you expect other people working for free to magically know your personal problems from your whining on slashdot?

Re:Experts...

[superwiz]

Just think about it! (TM) A function gets automatically executed just because you leave scope. Doesn't matter how you leave it. Forget that it's a destructor. It's a function which gets called automatically without anyone writing any code to call it. Show me how to do that in C. And that's the fundamental difference between the 2 languages. The rest is syntactic sugar.

Re:Simple ...

[serviscope_minor]

Congrats, you clearly have the bigger C++ e-penis in this case.

That's probably the most common type of logical phallusy.

Re:Experts...

[Smerta]

I've made the exact same argument to co-workers at many firms... namespaces (e.g. Timer_Init()), virtual functions (tables of function pointers), etc. can be approximated / kludged together... but automatically invoking a function at the right place (destructor and, let's face it, the constructor is pretty handy too) is something that has to be baked into the language, and C++ has it. I work in safety-critical systems, and knowing that I can't leave a function with interrupts disabled, I can't forget to close this socket, etc. is incredibly comforting.

I'll quote Bjarne Stroustrup here:

"Just that closing brace. Here is where all the ‘magic’ happens in C++. Variables get destroyed, memory gets released, locks get freed, files get closed, names from outside the closed scope regain their meaning, etc. This is where C++ most significantly differs from other languages. It is interesting to see how destructors -- an invention (together with constructors) from the first week or so of C++ -- have increased in importance over the years. So many of the modern and most effective C++ techniques critically depend on them"