The University where I studied didn't teach C++ to undergraduates, and as I understand it (I work there today, via a circuitous route, but not for the Computer Science department) it still doesn't. A friend teaches Haskell, and I believe their introduction course is in Java (mine was in ML).
I can't say that I'd think C++ would be a good choice. Perhaps someone who teaches C++ to undergraduates can offer their opinion about how that works out.
P2000 is an interesting document, and so thanks for causing me to read the current revision - but I'm not sure what you meant by "We need replacements". What is it that you want to replace? The C++ language? The C++ programmers?
It is clear that the way things are going, even the most passionate C++ developers can't make use of the language in a full safe way, understand the changes between revisions, and naturally Scott Meyers can't write books forever.
I certainly wouldn't manage a C++ interview quizz, even though I try to keep up with it.
The university where I studied during the 1990's, C was already on the way out.
First year students got Pascal and C++ with proper set of vector, string and other common data structures. This was a couple of years before C++98.
All data structures used bounds checking, and it was great, because the professor has a similar point of view as I, so C++ was taught as a kind of poor man's Ada in regards to safe coding.
What about C? Well, those taking OS classes were expected to be able to dive into it from C++ point of view.
However, this isn't normal, most institutes follow the more traditional path of C with classes, and most of them, hardly moved beyond C++98 anyway.
> even the most passionate X developers can't make use of the language in a full safe way
Is pretty much true for any generally used language X. You can write security bugs, invalid domain logic, code that kills people etc. in any language that’s popular today. However there are examples of proven correct software (CompCert, CertOS, seL4 etc.) so we will hopefully get there eventually.
I am pretty much in touch, contrary to you that argues for C++ no matter what, yet we seldom see any public work from you other than asserting how C++ is great and anything else is bad.
Apparently in your eagerness to once again argue for C++ in yet another C++ comment you translated "naturally Scott Meyers can't write books forever." as if I hasn't aware of the fact that he has left the community long time ago.
Apparently those words cannot have any other meaning as I am out of touch with what Scott Meyer does, otherwise how could you use it.
Anyone that bothers can assert many of my statements with public stuff from me on the Internet, they can even find naive Usenet stuff from me all the way back to 1994.
Prove us that your "expert C++ knowledge about the real market usage" is to be taken seriously and not polluting HN with C++ zealotry.
I learned it myself while at Uni because we only did C and Java.
I d say it s important to learn a bit just for the C with classes aspect. Sometimes you want to do a moderately complex C program that Java dev will have to help with, and it helps knowing C++ (as in: I did a 3D engine in it, using no overly fancy language feature, for me that s good enough to get an ability to produce something useful enough).
It's a bit sad Rust and Go are so far from C. Might as well just go to Java, a paradise for team work and problem solving.
I can't say that I'd think C++ would be a good choice. Perhaps someone who teaches C++ to undergraduates can offer their opinion about how that works out.
P2000 is an interesting document, and so thanks for causing me to read the current revision - but I'm not sure what you meant by "We need replacements". What is it that you want to replace? The C++ language? The C++ programmers?