Re: [boost] Review Wizard Status Report for September 2012

Subject: Re: [boost] Review Wizard Status Report for September 2012
From: Vadim Stadnik (vadimstdk_at_[hidden])
Date: 2012-09-29 11:57:37

On Sun, Sep 30, 2012 at 12:26 AM, Lars Viklund <zao_at_[hidden]> wrote:

> On Sat, Sep 29, 2012 at 07:43:32AM +1000, Vadim Stadnik wrote:
> > On Fri, Sep 28, 2012 at 11:22 PM, Lars Viklund <zao_at_[hidden]> wrote:
> >
> > >
> > > How does this library fare after the advent of Boost.Containers? From
> > > the description, it sounds like it overlaps quite a bit, to the extent
> > > that it can likely be merged with it.
> > >
> >
> > The library uses new data structures with improved computational
> complexity
> > of many algorithms against data structures used in C++ standard
> containers.
> > This is why this library supersedes facilities of standard containers. I
> > will add a section to the documentation with specific improvements.
>
> Very good, but I didn't ask about the standard containers, I asked about
> Boost.Container [1].
> Do yours overlap with his? Is your library really distinct enough to be
> a different library instead of merging with Ion's library?
>
> It just feels quite strange to have two libraries that both claim to be
> "better containers, lol".

 The previous version of stl_ext_adv was thoroughly reviewed by Joaquín M
López Muñoz. Boost does not have containers based on advanced data
structures, but your point is perfectly valid and the comparison with
Boost.Container should be added to the documentation. Also, I think that
merge of new and existing containers into one library is possible provided
that they support the same interfaces.

> > Additionally, surely there's a better name than "STL extensions"? First
> > > of all, some people (me) truly dislike the STL misnomer. Secondly, it
> > > doesn't convey no information about what it actually extends/provides.
> > >
> > > The full library name is “STL extensions based on advanced data
> structures“. I have shortened the name to “STL extensions” for the Review
> Schedule
> > only. It seems to me it is quite reasonable. For example, CGAL offers
> > Multiset container with similar advanced functionality and uses name “STL
> > Extensions for CGAL”. Suggestions are welcome.
>
> It's kind of hard naming what essentially seems to be a container
> library, considering that there already exists one. I still hold my
> opinion that "STL extensions for blargh" makes a really horrible Boost
> library name.
>
> These questions will come up constantly during the review, so you might
> as well get everything sorted out up-front.
>
>
I agree that right naming a library is a difficult task. Alternative names,
such as “Advanced containers” or even “Container++”, are possible. Let us
see what other experts think.