C++ in the modern world | Hacker News Reader

C++ in the modern world

medium.com

22 pointsfrostmatthew10 comments

The Qt Contributor's Summit just wrapped up today, and some of the discussions [0] touched on the author's points [1][2].

Unrelated to above, I just wanted to make a note: Just because the standard is out does not necessarily make the language easier to use; it is entirely compiler dependent. Also C++14 was not supposed to be as big of changes as C++11, if I recall correctly every-other standard is supposed to be major-minor changes to the language, though I don't have a citeable source on hand for that.

[0] http://wiki.qt.io/Qt_contributors_summit_2015_Program

[1] http://wiki.qt.io/QtCS2015_ModernCpp

[2] http://wiki.qt.io/QtCS2015_LocalQtMeetups

>though I don't have a citeable source on hand for that.

https://isocpp.org/std/status

In practice the major/minor thing hasn't worked out quite as intended. More stuff made it into C++14 than originally expected, and since they're prioritizing getting something out in 2017 it's currently looking like it'll be only a little bigger than C++14 (with C++20 possibly being significantly larger). Personally I'm glad that they're trying to stick to regular releases with whatever's ready.

More often than not I've read this almost-mantra 'use of C++ is declining' or 'C++ will be dead soon' or similar in threads on HN. A couple of times I chimed in and asked the authors of these statements for any proper sources. Never got an answer, but now it seems this TIOBE index could be used to actually even counter those statements immediately with what seems like solid evidence. However I am not too familiar with TIOBE; is it generally speaking regarded as an authorative source? Are there similar sources like this?

Tiobe is a joke. It is good enough for a very tough idea, but that is it.

That said, no programming language ever dies. Things are still being written in Cobol and common lisp. This doesn't mean focusing on them is likely to get you an interesting career with lots of choices.

HN is a bad source to use if you want to see what programming languages are used or not. One can find out perhaps which languages are fashionable or popular with certain types of companies.

C++ is not going to die until there is a much better alternative to it. Currently, there is none, and so C++ is in no danger. It has it's niche and it is not going to go away any time soon.

C++ is a legacy language like PL/I. Those who don't use it yet will not even consider it for new projects.

Certainly it will be considered if your project has certain kinds of requirements other languages fails to cover (no vm, highest performance, no gc, ...). Those are very relevant to a lot of projects. C++ is not a legacy language, it's sometimes (with C) the only viable language in some sets of conditions.

Funny how many languages' runtime on the TIOBE top ten are written in either C++ or C. Legacy indeed!

Reply on news.ycombinator.com