Yeah, and that reply is from CEO of Epic Games. I don't know if this holds any weight with ISO, but it at least explains why this particular Twitter thread complaining about access to ISO-8601 standard is hitting HN frontpage.
And he raises a good point downthread, too:
"So, what's the calculus with something like the ISO C++ standard? You sell 1000 copies at $200 = $200,000 lifetime income. Meanwhile, a lack of understanding of C++ details is kneecapping a hundred billion dollar segment of the technology industry. It's absolute madness."
From what I've heard most compiler developers base their work off those final drafts, so you'd be better off using them anyway. Might be an old wives' tale though.
I work at a compilers company and my manager has told us to not work off the drafts and to expense the standard if we lack a copy.
I suspect volunteers working on floss compilers use the draft, but anyone who can expense it has no reason not to.
As an aside, when I need to explain how something works for C, I quote the standard. When I need to explain how something works for C++, I quote Stroustrup's book. The C++ standard is very precise, but also very math-y. Stroustrup is much more approachable.
> I suspect volunteers working on floss compilers use the draft, but anyone who can expense it has no reason not to
This is the best summary of why having to pay for standards is a problem. Let's unpack it carefully.
> volunteers working on floss
An enormous, perhaps the most vital, part of the people working on everyday software stuff.
> anyone who can expense it
Companies and organizations that have money and may or may not have a de facto controlling market share of software that implements a spec
What happens when that company with a controlling market share implements a part of the adopted standard that differs in some apparently insignificant way from from the freely available draft? If that difference turns out to be more significant than it appears, now that company has the only truly spec-compliant implementation, and the rest of the market can't interop because of that difference.
> Meanwhile, a lack of understanding of C++ details is kneecapping a hundred billion dollar segment of the technology industry. It's absolute madness.
I agree with the sentiment that a lack of knowledge is kneecapping C++ development.
However, that lack of knowledge isn’t so much about the details, but the big picture. Forget reading the ISO C++ standard, I think most C++ developers would be better served by reading Scott Meyers “Effective C++” series. Forget about knowing the grimy details of of the standard. I would be happy for them know RAII, and avoid naked ‘delete’ calls.
I have some decent C++ experience, but even I try to avoid areas of the language where you have to be a language lawyer trying to parse if what you are doing is undefined behavior.
Yes, ISO c++ standard is the foundation, but there are a lot more idioms and other knowledge to learn about C++ that is far more important for making good C++ developers.
The target of the ISO standard is not your average developer, but rather the people who write the compilers, standard libraries, and books that bring C++ tools and knowledge to the general developer. For them $200 is nothing compared to the time and effort of those endeavors.
I don't think you can really understand C++ by reading the standard to be honest. If you're such an expert you can read the standard in its original form, you're probably already a seasoned developer anyway.
Usually when you're reading a standard like that it's because you're looking for details from an implementer's perspective, not the user's. The specifications have to make everything including the obvious details explicit, which leads to a lot of verbal abstraction that's counterproductive to just learning how to use the thing.
Once you're actually trying to implement the system though, suddenly all of that language becomes really helpful.
I doubt that "learn how to speak C++" is why anyone _does_ read the standard. The standard is used for things like ensuring broad compatibility among compiler implementations.
CLARIFICATION: I suppose that might have been your point--that not much hindering of understanding is going on among the broad C++ community due to standards (un-)availability.
Yeah exactly, I'd even go as far as saying that the standard is only useful if you're implementing a compiler or tracking down a bug in an existing one.
I'm realizing now that the original question was more about those developers doing work on C++ compilers and such than the broad community. Even so, I guess every heavy contributor and project like GCC, LLVM... have bought their own copy of the standard.
And he raises a good point downthread, too:
"So, what's the calculus with something like the ISO C++ standard? You sell 1000 copies at $200 = $200,000 lifetime income. Meanwhile, a lack of understanding of C++ details is kneecapping a hundred billion dollar segment of the technology industry. It's absolute madness."