Boost logo

Boost :

Subject: Re: [boost] Graph Algorithms
From: Siddhartha Sen (siddhartha99.sen_at_[hidden])
Date: 2019-02-22 13:58:05


Thanks for the feedback! But what if I focus on C++14 then? Is that a good
idea?

On Fri, 22 Feb, 2019, 6:14 PM Vinnie Falco, <vinnie.falco_at_[hidden]> wrote:

> On Fri, Feb 22, 2019 at 12:49 AM Siddhartha Sen
> <siddhartha99.sen_at_[hidden]> wrote:
> > From what I've read in the mailing lists, I think the demand has
> shifted
> > from C++11 to more modern versions, like C++14 and C++17. Anyways,
> > thanks for the tip. Will you be mentoring this year?
>
> I will not be mentoring directly although I do offer advice on the mailing
> list.
>
> People say they want C++14 and C++17 but I don't think its a good idea
> to require those language versions in a library, because most people
> are still on C++11. Have a look at the latest developer survey:
>
> <https://www.jetbrains.com/research/devecosystem-2018/cpp/>
>
> If a library requires C++17 then most C++ users will not be able to
> use it. And this is bad for the library, because fewer users means
> less testing, less feedback, less popularity, and in general this
> makes the library not as good.
>
> In terms of the API (interface) of a library, C++17 does not offer
> much more than C++11 except for things like, avalability of
> string_view, variant. But Boost has these things already (e.g.
> boost::string_view, boost::variant) and they work in C++11. So there
> is really no benefit to C++17 other than some improvements in the way
> you can write the implementation. But that doesn't affect users
> directly.
>
> In summary, by requiring C++17 most users will be unable to use the
> code, and there will be very little benefit to C++17 users.
>
> Regards
>


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk