It's surprising that C++'s development trend continues.
When a game or program is made with C++, it's usually nice because performance is mostly guaranteed. But if someone told me to write C++ myself, I'd cry. There's too much to memorize, and the standards are too varied. When I go to a project site for maintenance and it's a C++ project, I instantly lose energy — because it's just too difficult.
I'd be happy if someone else wrote it, but it's not a language I want to write myself
Personally I don't find programming with C++ that hard. The downside is it needs a brain warm-up, and this is per project, but once that flywheel is spinning, I find it almost effortless to write code.
I have to go through the same warm-up more or less for any language I work with, so it's not that different than writing Python, Go or Java for me.
The difference, if you split hairs, that the brain warm-up takes a bit longer. Maybe a couple of hours, or a day at most.
Otherwise it's not different for me. I don't feel different while writing with any other language. I guess the main reason is I always think like the computer first and translate that thinking to the programming language at hand.
There are so many standards and idioms that it gets confusing. There are still legacy codebases out there — some codebase still use C++98 as their standard, others use C++11... And with Unreal Engine, the modern C++ standard is C++14, right? There are things like smart pointers, but some places don't even use them. I feel like there are just too many features. When I saw template metaprogramming — that new feature — I realized I have no talent for C++.
I have developed things with C++98, C++11 and C++14. Every of these standards are so vast, so remembering everything (even in a single standard) is not possible. Instead of knowing everything, I first fix the standard I want or need to work with.
Then I design the thing I want to build. I always design what I want to build beforehand. This takes a couple of iterations from high level to low-ish level. That last design becomes a bit language dependent. Then I select some of the core tools that I'm going to use (which kind of pointers, classes or structs, etc.)
With that design in mind, I go "library shopping" both for file formats (if any) or other stuff like vectors, etc.
Armed with the reference docs of these, I write my code with the toolbelt I have built for the project.
Some things are hard, but they are not impossible. I find thinking like compiler helps a lot.
This is true of any language. Python with flask vs django, with/without type hints. JavaScript with anhular and vue.
The varying standards are no different to major python versions or go versions - arguably there’s even less between most versions than there is in your average go release.
The differences in apps and frameworks don’t matter for day to day - std::string, Unreal’s FString and QT’s QString all are similar enough that 99.9% of the time.
Metaprogramming is one of those things; you either write it or you don’t. Knowing some basics is required but the vast majority of people use a handful of pre existing things without understanding the nuances of how it works under the hood.
> When I saw template metaprogramming — that new feature — I realized I have no talent for C++.
It's not a new feature. And tbh, compared to Typescript, C++ templates are tame ;)
(but yeah, deciding when to stop digging into the template metaprogramming rabbit hole requires some common sense and sanity, too much template complexity is almost never worth the hassle)
You're right that C++ has a lot of features. But like mentioned elsewhere most projects define their own conventions and the subset of features that they use.
Also the nice thing about having a large set of features is that C,++ allows you to write very nice abstractions (or not) at both very low or at very high level. In other words you can be very low level with online ASM and bit operations and bit and direct memory manipulation or very high level almost like a script language. Whatever the problem domain needs C++ has got you covered.
For games, C++ becomes a much simpler language since game code bases usually ignore the C++ stdlib (at least mostly, and for good reasons, e.g. see [0]). And without the stdlib C++ is actually kinda-sorta okay-ish.
Related, the main problem with the C++ ecosystem is that everybody carves out their own language subset, so it's not one ecosystem but many ecosystems with contradicting styles and language/stdlib subsets. This makes code reuse via libraries much harder than it should be.
I fully agree. In my personal project, I ended up using the STL to get off the ground, but in the end I replaced pretty much everything with custom-written code.
Once you get rid of the STL, compile times get so much better. With modern c++23 features, templates actually become really convenient to write, and at the core there is a really useful and pleasant to use language.
I try to avoid c++ libraries and instead rely on c-style APIs. Usually the c++ style libraries force you into using the STL, which comes with a heavy tax on compile times, without much benefit in comfort of use.
On the contrary. You can focus exactly on the features the higher level game code needs. The C++ stdlib is (for the most part) poorly designed, usually poorly implemented, the main reason for slow build times, and its complexity explodes because it needs to consider all edge cases that most code bases don't ever trigger.
A specialized dynamic array class in a few hundred lines (at most!) and with just the required features is much more useful than the 20kloc monster that's pulled in with `#include <vector>` and which doesn't even do bounds checking in the 'idiomatic' usage.
Saying it doesn't even do bounds checking (in release builds) is to miss one of the major points of C++ - not paying for what you don't need. It's not a mistake, it's a feature.
You complain about it not being suitable for game development in one comment but then expect bounds checking in release builds? You're sitting in multiple lanes at the same time.
NIH implementations are usually grossly inferior because as it turns out, it's quite hard to get it right and those edge-cases aren't important until you start getting bitten by them when you'd rather be shipping features.
Bounds checking overhead is negligible for all but the absolutely hottest code paths (fwiw we shipped active asserts, including bounds checking asserts in all the PC games I was involved with - carefully monitoring the overhead of course).
The main reason to not use the stdlib isn't so much about squeezing out the last bit of performance, but about control of what actually happens under the hood (and also compilation times). The overall runtime cost of all those active asserts (not just the range checks, everything) was somewhere in the 2..3% range, which is fine when budgeted for upfront.
I find it hard to agree that the stdlib is poorly designed and implemented. In my entire career it has pretty much worked entirely to spec.
Yes, it can exhibit non-optimal performance, and in some specific cases (regex's especially), extremely poor performance, but that's not the same as being poorly designed and implemented, especially given the breadth of the thing.
Yes I don't disagree that sometimes a specific container or a data structure is great for the problem. Problem is that most of the game code and related code (so tooling,editor, auxiliary engine code) does need a typical STL type functionality and then when the org has "omg no STL" blanket rule someone ends up implementing STL and that's almost always worse than the STL that ships with the tool chain. Even worse..it'll be missing features and data structures and then people have to write sub-optimal code to work around it's limitations.
some of the STL is easy to improve on. For example, std::unordered_map performs poorly due to pointer stability requirements in the standard. Most performance sensitive C++ codebases will use something like abseil's hash maps instead.
Grossly exaggerated or misunderstood in many cases.
Some of their arguments are just flat-out wrong.
I mean, why are they blaming the standard library for inherent properties of linked lists? Yeah, you don't want to use them without good reason. That's just called picking the right data structure for the job, not a flaw with the standard library.
Some of the other choices were tradeoffs between performance and usability. The standard maps have stable iterators, whereas third-party implementations almost never do because you can write faster implementations if you're willing to live without those guarantees. Was it the right choice in hindsight? Maybe, maybe not.
I'd personally like to see a namespaced versioned standard library but like that's ever going to happen
As I understood the article, the main critique is that the stdlib has no concept of deprecation and breaking backward compatibility. E.g. the C++ committee is quick to add badly designed features to the stdlib but then can't roll them back when people actually realize that those new features are useless for most real-world code.
You can be pretty productive even with 70% of the language :) It is a common misconception that C++ is suitable only for game engines and similar domains. It is perfectly fine for applications domain as well.
As a side note, regarding your profile info, unless you are based in North Korea, please at least add one 0 to your rate. You'll get more long-term and high-quality clientele.
> You can be pretty productive even with 70% of the language
Or even far less than that. I like to use it as C with lambdas and namespaces. Sprinkle in metaprogramming as needed. Even just not having to remember to call cleanup code thanks to dtors would alone be enough to sell me on it.
Honestly, I don't expect to find clients here. Fundamentally, you have to trust me to give me work. The amount of money doesn't really matter much to me.
The ecosystem isn't fine - just to get a project going requires picking a non-trivial set of tools and approaches, none of which the C++ standard enforces or guides to.
For example, will you manage dependencies via packages? If so, with what? What will you use for building your project? The list goes on and on.
- new features overlapping old features from previous standards without replacing them or deprecating them (function::copyable_function vs std::function, std::less<> key for transparent lookup in maps)
- new features not usable by the layman (coroutines ...)
- Cryptic syntax (reflection...)
- Stuff you are told not to use because of performance reason and that cant be fixed because of ABI (regex)
- Compile errors that are 1km long (no, concepts are not helping here, the 'nicer' message is still buried into a hot pile of template instantiation callstack).
I personally find the lack of native package management in C++ as a blessing. Go, Python, Rust has it, and this always causes pulling in infinite number of packages for any trivial operation.
sudo-rs was pulling in 1M+ LOC as its dependency chain at one point. I believe they removed the biggest offenders, but I didn't check it recently.
>There's too much to memorize, and the standards are too varied. When I go to a project site for maintenance and it's a C++ project, I instantly lose energy — because it's just too difficult.
If you'd already been using it for 10+ years you wouldn't feel that way, because you'd already have memorized a lot of it.
You don't have to use them. There's a handful of nice to haves in modern releases but its totally fine and sane to just ignore whatever the committee is distracted by at the moment.
Hell, if you wait long enough, they'll just deprecate it before you can care to bother.
Since I've been working in C++ a lot recently I decided to watch the video as I waited for a build to complete. So the length is about right. And fortunately, the video is a delight!
So happy to see Andrei Alexandrescu was included in this documentary. His book on modern C++ design was a mind opener at the time I read it. Maybe still is today. Anybody else read it?
Yes, I feel the same way. I met Andrei once on a Meetup in Munich, basically telling him that he taught me how to think which led to a somewhat awkward conversion. Fun times nonetheless :-)
I‘m out of the loop: we‘ve had Python, Clojure and possibly something else recently. Is that a series by the same people working through several languages? Is it happenstance? Is it a trend, and every programming language is now scrambling to get their own video documentary?
My only problem with C++ is that it’s too verbose. my eyes need to parse huuge chunks of things when I just want some convenient syntax for it. otherwise the idioms are pretty universal for most programming languages nowadays.
A few reasons:
1. Header files make C++ verbose. Header files are well within LLM's ability
2. LLMs can handle setting up cmake for you
3. C++ is very well documented relative to (most) newer languages
4. LLMs can port modern features like websockets and build API wrappers easily, reducing the disadvantage against web (since most documentation is for JS/python/go)
Coding languages have been developing for speed of (manual) writing - akin to how human languages did with modern alphabets. Now that writing is a lot easier, languages will likely evolve towards a focus on execution (or in the case of human languages, speed of reading and precision of understanding)
Herb's blog post links to the SlashData Developer Nation Survey, so presumably that's what the claim is based on. The company has a methodology page here [1], and it looks like the Developer Nation panel [2] is one of the sources used by that company.
I have the utmost respect for Casey, but his disdain for Stroustrup is unfounded. The fact of the matter is C++ occupied a niche in the right place and at the right time, and it grew from there. Many mistakes have been made, but Stroustrup is in no way personally responsible for all of them and I don't think Stroustrup is a bad programmer (something I've heard Casey say in some of his videos). You can argue that the committee route is not the best, but C++ is here to stay and by some metrics adoption is actually growing.
Strangely, I've never seen any nice code from Casey despite all of the mud slinging he's done over the years. Maybe it exists somewhere but I watched a lot of Handmade Hero when it was starting off and the code was a mess.
It feels as though he just attracted an audience of junior developers who take everything he says as gospel, as is often the case with social media programmers. Lord knows I've argued with some of them and they usually crumble as soon as they don't have one of his opinions to throw back at you.
I agree. His negativity has probably detracted quite a few people from him that otherwise are quite aligned. Still, his historical remark is rather peculiar.
As for the language, yes, sadly, it’s with us seemingly to stay. I code it professionally and I can’t find a single interesting, or even good, thing about it. Apart from wide adoption of course. Everything about it feels extremely badly designed from the user perspective (though it’s probably technically very impressive) with many details, that probably the sanest strategy is to use a small subset of the language. At least I don’t have to use STL at work, that’s something positive, I guess :)
Casey also got less aggressive when talking about Stroupstrup lately, especially after his last talk at Better software conference, where he mentioned him multiple times with a lot more historical context.
When a game or program is made with C++, it's usually nice because performance is mostly guaranteed. But if someone told me to write C++ myself, I'd cry. There's too much to memorize, and the standards are too varied. When I go to a project site for maintenance and it's a C++ project, I instantly lose energy — because it's just too difficult.
I'd be happy if someone else wrote it, but it's not a language I want to write myself
I have to go through the same warm-up more or less for any language I work with, so it's not that different than writing Python, Go or Java for me.
How is that different from other languages, which don't need the brain warm-up?
Otherwise it's not different for me. I don't feel different while writing with any other language. I guess the main reason is I always think like the computer first and translate that thinking to the programming language at hand.
Then I design the thing I want to build. I always design what I want to build beforehand. This takes a couple of iterations from high level to low-ish level. That last design becomes a bit language dependent. Then I select some of the core tools that I'm going to use (which kind of pointers, classes or structs, etc.)
With that design in mind, I go "library shopping" both for file formats (if any) or other stuff like vectors, etc.
Armed with the reference docs of these, I write my code with the toolbelt I have built for the project.
Some things are hard, but they are not impossible. I find thinking like compiler helps a lot.
The varying standards are no different to major python versions or go versions - arguably there’s even less between most versions than there is in your average go release.
The differences in apps and frameworks don’t matter for day to day - std::string, Unreal’s FString and QT’s QString all are similar enough that 99.9% of the time.
Metaprogramming is one of those things; you either write it or you don’t. Knowing some basics is required but the vast majority of people use a handful of pre existing things without understanding the nuances of how it works under the hood.
It's not a new feature. And tbh, compared to Typescript, C++ templates are tame ;)
(but yeah, deciding when to stop digging into the template metaprogramming rabbit hole requires some common sense and sanity, too much template complexity is almost never worth the hassle)
Also the nice thing about having a large set of features is that C,++ allows you to write very nice abstractions (or not) at both very low or at very high level. In other words you can be very low level with online ASM and bit operations and bit and direct memory manipulation or very high level almost like a script language. Whatever the problem domain needs C++ has got you covered.
Related, the main problem with the C++ ecosystem is that everybody carves out their own language subset, so it's not one ecosystem but many ecosystems with contradicting styles and language/stdlib subsets. This makes code reuse via libraries much harder than it should be.
[0] https://hftuniversity.com/post/the-c-standard-library-has-be...
Once you get rid of the STL, compile times get so much better. With modern c++23 features, templates actually become really convenient to write, and at the core there is a really useful and pleasant to use language.
I try to avoid c++ libraries and instead rely on c-style APIs. Usually the c++ style libraries force you into using the STL, which comes with a heavy tax on compile times, without much benefit in comfort of use.
On the contrary. You can focus exactly on the features the higher level game code needs. The C++ stdlib is (for the most part) poorly designed, usually poorly implemented, the main reason for slow build times, and its complexity explodes because it needs to consider all edge cases that most code bases don't ever trigger.
A specialized dynamic array class in a few hundred lines (at most!) and with just the required features is much more useful than the 20kloc monster that's pulled in with `#include <vector>` and which doesn't even do bounds checking in the 'idiomatic' usage.
You complain about it not being suitable for game development in one comment but then expect bounds checking in release builds? You're sitting in multiple lanes at the same time.
NIH implementations are usually grossly inferior because as it turns out, it's quite hard to get it right and those edge-cases aren't important until you start getting bitten by them when you'd rather be shipping features.
Bounds checking overhead is negligible for all but the absolutely hottest code paths (fwiw we shipped active asserts, including bounds checking asserts in all the PC games I was involved with - carefully monitoring the overhead of course).
The main reason to not use the stdlib isn't so much about squeezing out the last bit of performance, but about control of what actually happens under the hood (and also compilation times). The overall runtime cost of all those active asserts (not just the range checks, everything) was somewhere in the 2..3% range, which is fine when budgeted for upfront.
Yes, it can exhibit non-optimal performance, and in some specific cases (regex's especially), extremely poor performance, but that's not the same as being poorly designed and implemented, especially given the breadth of the thing.
https://en.wikipedia.org/wiki/Alexander_Stepanov
The standard library is mostly fine to use unless you have specific needs.
The bit about libraries is nonsense, sorry.
I mean, why are they blaming the standard library for inherent properties of linked lists? Yeah, you don't want to use them without good reason. That's just called picking the right data structure for the job, not a flaw with the standard library.
Some of the other choices were tradeoffs between performance and usability. The standard maps have stable iterators, whereas third-party implementations almost never do because you can write faster implementations if you're willing to live without those guarantees. Was it the right choice in hindsight? Maybe, maybe not.
I'd personally like to see a namespaced versioned standard library but like that's ever going to happen
As a side note, regarding your profile info, unless you are based in North Korea, please at least add one 0 to your rate. You'll get more long-term and high-quality clientele.
Or even far less than that. I like to use it as C with lambdas and namespaces. Sprinkle in metaprogramming as needed. Even just not having to remember to call cleanup code thanks to dtors would alone be enough to sell me on it.
The ecosystem isn't fine - just to get a project going requires picking a non-trivial set of tools and approaches, none of which the C++ standard enforces or guides to.
For example, will you manage dependencies via packages? If so, with what? What will you use for building your project? The list goes on and on.
The language keeps growing, with
- new features overlapping old features from previous standards without replacing them or deprecating them (function::copyable_function vs std::function, std::less<> key for transparent lookup in maps)
- new features not usable by the layman (coroutines ...)
- Cryptic syntax (reflection...)
- Stuff you are told not to use because of performance reason and that cant be fixed because of ABI (regex)
- Compile errors that are 1km long (no, concepts are not helping here, the 'nicer' message is still buried into a hot pile of template instantiation callstack).
sudo-rs was pulling in 1M+ LOC as its dependency chain at one point. I believe they removed the biggest offenders, but I didn't check it recently.
If you'd already been using it for 10+ years you wouldn't feel that way, because you'd already have memorized a lot of it.
- new features overlapping old features previous standards without replacing them or deprecating them. - new features not usable by the layman - ...
See function::copyable_function vs std::function, modules, coroutines, Reflection syntax is cryptic at best, ...
Hell, if you wait long enough, they'll just deprecate it before you can care to bother.
Because of AI, right?
Coding languages have been developing for speed of (manual) writing - akin to how human languages did with modern alphabets. Now that writing is a lot easier, languages will likely evolve towards a focus on execution (or in the case of human languages, speed of reading and precision of understanding)
Is this good or bad?
(if you mean that film, most likely no.)
[0]: https://www.slashdata.co/research/developer-population
[1]: https://www.slashdata.co/company/methodology
[2]: https://developernation.net/
It feels as though he just attracted an audience of junior developers who take everything he says as gospel, as is often the case with social media programmers. Lord knows I've argued with some of them and they usually crumble as soon as they don't have one of his opinions to throw back at you.
As for the language, yes, sadly, it’s with us seemingly to stay. I code it professionally and I can’t find a single interesting, or even good, thing about it. Apart from wide adoption of course. Everything about it feels extremely badly designed from the user perspective (though it’s probably technically very impressive) with many details, that probably the sanest strategy is to use a small subset of the language. At least I don’t have to use STL at work, that’s something positive, I guess :)
You can argue that chemical companies route is not the best, but cancer is here to stay and by some metrics adoption is actually growing.