Hacker News Reader

Comment by MathMonkeyMan | Hacker News Reader

The C/C++ people hate it when you call it C/C++.

That's a fairly recent thing that some people do as a kind of virtue signaling.

The "C/C++ Users Journal" was a very popular publication and no one took issue with its name. Nor do people take issue with Dr. Dobbs which has a "C/C++" section with articles from highly influential members of the C/C++ community.

C++ is a complex language, so people invent ways to show how dedicated they are to it, and nowadays that means pretending to get all worked up over the term "C/C++".

I think it's also that idiomatic C++ is evolving further and further away from anything resembling C. So it's turning from C with extra features into a completely separate language that happens to interoperate well with C.

This is true. It's also true that, in terms of abstract machine semantics, C++ is the closest language to C. Similarly, in terms of optimizing compiler implementations, they're the two "core" languages implemented by most frontends. The grouping of "C/C++" makes sense in both of these contexts, at the minimum.

C/C++ is undefined if C is an instance of an integral type.

it's just 1/++

C/C++ is as naïve as ASM/C or Ape/Human.

You come up with examples that are over 30 years old.

Yes my examples are old because the entire point I'm making is that this outrage over "C/C++" is a relatively recent phenomenon. People 10-20 years ago didn't really care or take issue with that term.

I was there and yes we did ;)

I was there and yes we did.

You wish it was recent and something you could pretend about. But anytime you write it, you are telling a lie. Being seen lying tells the reader more about you than you probably intend.

> That's a fairly recent thing that some people do as a kind of virtue signaling.

If by recent you mean well over 20 years ago. When I began learning C circa 2000, I used to hang out on comp.lang.c, and "C/C++" was very much frowned upon. And for good reason--people felt entitled to ask C++ questions in a C newsgroup, which was and remains a poor idea. Even if they were asking questions relevant to both, e.g. regarding pointers, it's still a much better idea to ask in a C++ newsgroup. See https://en.wikipedia.org/wiki/XY_problem

There are people (invariably young) who resist newsgroup etiquette, convinced the etiquette is pointless, or something like virtue signaling. Over the decades they eventually come around to appreciating the wisdom, even if there's never a mea culpa.

The commercial software world, and especially the Windows world, often had a much different culture than what developed online, especially in the FOSS and Unix subcultures. For the latter especially, C++ was relatively uncommon until the 21st century. C++ (with KDE, etc) and then Python had the effect of importing or creating whole new subcultures in these communities without any memory or appreciation of the pre-existing culture.

Not seeing evidence of your claim. A simple search on comp.lang.c for C/C++ turns up plenty of results and no one, not a single person, is calling it out. Infact there's plenty of discussion on an almost daily basis discussing the two languages almost seamlessly. Looks like there's even a alt.comp.lang.learn.c-c++ newsgroup as well.

For reference I also did a search from 1998 to 2002... once again no one cares or is objecting to the term C/C++ and that term is used quite often. From this search I also just learned that Microsoft's own compiler uses the term "C/C++".

Furthermore, if something is relevant to both languages, such as a question involving pointers, why would it be a bad idea to ask about it in a C newsgroup? If anything, it's preferable to ask a question that is pertinent to both languages, especially one involving pointers, in the C newsgroup.

Certainly asking a strictly C++ question in a C newsgroup would be off-topic and rightly so, but asking a question about pointers or other low-level details that are applicable to both languages is perfectly fine in a C newsgroup.

Using 'group:comp.lang.c "C/C++"', here's one of the first hits[1] I get, from May 17, 2000 and starting off,

> I'm just curious, but why is it that people in this group tend to reply with "the language C/C++ does not exist" or "C/C++ is 2, what do you mean?"

https://groups.google.com/g/comp.lang.c/c/bVBTSJ3KRrg/m/uv98... [2]

followed by 14 responses in that thread, most of which explain why that is, none defending the usage. One of the respondents is active on HN, I believe: kazinator (Kaz Kylheku).

I have a local copy of the UTZOO archives, with posts from the 1980s. Out of ~30,000 posts in comp.lang.c and comp.std.c only 284 use the phrase "C/C++". Many of those are discussing specific compilers, e.g., "Zortech C/C++" (this is especially true in other newsgroups). In a quick survey I couldn't find one where people corrected the usage, but if you read those threads most people are careful to provide answers specific to C and don't often repeat "C/C++". Presumably after C and then C++ became standardized people became more vocal about treating them as distinct.

[1] Results weren't sorted by date, and the first page of results spans 30 years.

[2] I wish I could also include a reference ID from the headers, but Google Groups apparently no longer makes them accessible even when logged in. Google Groups Usenet support is horrible, even when searching. It's a real shame. I think Google Groups has even blacklisted comp.lang.c for spam, though almost all the spam in comp.lang.c comes from Google Groups. I've filtered out Google Groups posts from my feed for the past 10 years or so.

Here is a thread starting on April 15, 2000:

https://www.usenetarchives.com/view.php?id=comp.lang.c&mid=P...

started by someone advertising a C/C++ tutorial.

Here, down in the replies, we can find the remark:

In the first place, there is no such language as "C/C++",

Another one:

BTW, what in the world is the C/C++ language

And:

First problem: is comp.lang.c/c++ a legal name for a Usenet forum ?

And:

  Jones
  16/04/2000 17:40:46 UTC

  mike burrell wrote:

  > x*....@m*-****a.com wrote:

  > > I'm beginning a new C/C++ tutorial over at

  >

  > sweet. it's the c/c++ thread all over again.

  Not from me! I'm keeping my mouth shut this time ;)

  Indi

Look, just believe the people who were there. I was; my name appears in that thread.

Another thread, same year, "NEED C/C++ PROGRAMMER", started by "ITJOBS":

https://www.usenetarchives.com/view.php?id=comp.lang.c&mid=P...

Remarks:

"No such language as C/C++."

"Presumably ITJOBS thinks "Well, I've heard of C and I've heard of C++, but I don't know anything about either. They should be about the same."

"I see this a lot and just want to add that a slash ("/") does not imply a mixture of the two, it just means "and/or", so it would read "C and/or C++".

"All the left-brained, analytical people on comp.lang.c hate "C/C++". The right-brained, "creative" people don't mind it. End of story."

"C/C++ is 1 unless C is 0 in which case the result is undefined (I couldn't resist)"

"Um, no. It's always undefined."

> A simple search on comp.lang.c

I'd love to know where such a thing is possible today.

The distinction actually makes perfect sense during the hiring process.

If a recruiter approaches me with a C/C++ gig then I will immediately assume that either the hiring manager doesn't know what they want or the hiring manager didn't actually write the job specs, none of those scenarios sends a positive signal.

> That's a fairly recent thing that some people do as a kind of virtue signaling.

It was a pretty common correction the comp.lang.c newsgroup in the 1990s.

"Hey you dolt, there is no C/C++; they are different languages."

"Oh yeah? Well look at how when I run cl.exe, it says Microsoft C/C++ compiler!"

"Haha, C/C++ is actually undefined behavior: C++ is modified and independently accessed in the same expression ..."

Fair enough. I'll stick to C and C++/Rust.

You see it a lot more often from the Microsoft world since MSVC only supported "C/C++" as a combined language, there was no separate C compiler.

They had a C compiler for kernel drivers.

Really? I don't know much about microsoft's internal architecture after Windows 98. But they maintained a separate compiler for what was basically just kernel modules?

It was a command line only tool distributed with the driver dev kit. It had ABI differences with MSVC that required them to keep it around until relatively recently.

MSVC is a C++ compiler only. It doesn't actually support C.

It does support C, most recently the C11 and C17 ISO standards.

https://devblogs.microsoft.com/cppblog/c11-and-c17-standard-...

It does, but I know why you have this interpretation and I think it's fair to have it.

MS only created their compiler to support C all the way up to C89 and it's original standard, with some Microsoftisms in it. MS did not update their compiler to support C99 and subsequent updates until, I think the late 2010s. Since then, they have updated and supported some of the newest C standards.

Well, technically it is one compiler, but it can operate in different modes, depending on the file input. So I don't this distinction is all too relevant.

“It’s time to halt starting any new projects in C or C++”

I personally don't see this level of pedantry from the people I work with. Even if C90/C99/C11/C++98/C++11/C++17/etc are all different, we understand people use "C/C++" to mean "that pervasive family of languages which most systems and libraries are built on".

I certainly don't. Although I also write Rust now, so I suspect someone will accuse me of having split loyalties or something equally ridiculous.

The only context in which I care about "C/C++" is on resumes, where it's a sign that I need to ask a little more about the person's background.

> The only context in which I care about "C/C++" is on resumes, where it's a sign that I need to ask a little more about the person's background.

Is that because you'd expect anyone actually experienced in either to know better than to lump them together so casually?

Not necessarily -- I've interviewed plenty of people where "C/C++" meant "experienced in both, but most work was in C with a small smattering of C++" (or vice versa).

In my experience, it breaks roughly between half that and half recent graduates who took one C course and one C++ course.

For me, I see “C/C++” as a sign they see C++ (which I’m looking for experience in) as interchangeable with C. I want people who grok C++’s ability to build powerful abstractions, not people messing around with it as C with polymorphic classes.

Yes. "C/C++" on a résumé is a red flag.

"Next!"

It's a sign of a liar, and you can round-file it without reading any more.

So far that on a CV has always meant "can write neither C nor C++"

The correct name is GNU Visual Clang++

It's GNU/C/C++ !!!

Reply on news.ycombinator.com