Boost logo

Boost :

Subject: Re: [boost] [GSoC] [Boost.Hana] Formal review request
From: Louis Dionne (ldionne.2_at_[hidden])
Date: 2014-07-30 18:25:10


Mostafa <mostafa_working_away <at> yahoo.com> writes:

>
> On Wed, 30 Jul 2014 13:19:44 -0700, Louis Dionne <ldionne.2 <at> gmail.com>
> wrote:
>
> [...]
>
> If you haven't/aren't already doing so, you can explicitly document:
> 1) That this is an FP library.
> 2) Why it is so and how this impacts the naming and design. Explicitly,
> why a non-FP approach would make for a sub-par library.
> 3) When any unfamiliar (to C++ programmers) names are first introduced
> provide (the approximate?) C++ mapping.

Hmmm. Yes, it definitely makes sense to at least put a warning sign and
to give rationales. That would basically do the job I'm currently doing
on this list, but once and for all. Will do.

> A part of the documentation may just be about educating a FP-illiterate
> audience. It would also be help if some of the C++14 concepts were
> explained, or in the least their use highlighted with appropriate links to
> references.

I won't write a tutorial on C++14, but I'll put relevant links when I use
something that's new in the language. Thanks for the suggestion, I had'nt
thought of that.

> One last thought, as Joe-everyday programmer why should I care about this
> library? What does it allow me to do (or do easier) that I wasn't able to
> do before?

If you either use MPL/Fusion or do any non-trivial std::tuple manipulation,
Hana will reduce the complexity of your metaprogramming code by giving you
high level abstractions at low compile-time cost. For example, instead of
doing like those folks did here
    
    http://stackoverflow.com/a/20441189/627587

just to unpack a std::tuple into a variadic function, you would write
    
    unpack(f, std::make_tuple(...)) == f(...)

with Hana. And it's going to compile just as fast as the hand-written
version. If you're willing to use the tuple provided by Hana, it's going
to compile much, much faster.

If you don't do any kind of metaprogramming, Hana is not for you.
  


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