Re: [boost] radix sort

Subject: Re: [boost] radix sort
From: Jeremy Murphy (jeremy.william.murphy_at_[hidden])
Date: 2013-07-21 08:29:40

Just to be clear, Steven's code is not actually radix sort, it's bucket
sort, right?  Which is great, but let's not get confused about what's what.

Steven, I gave it a quick test but it failed when trying to include
<boost/static_warning.hpp>.  Looks like Robert Ramey changed the names of
things around 1.37.0.  (I tested with 1.53.0.)  Changing it
to <boost/serialization/static_warning.hpp> fixed it up.

Looks like <vector>, <cstring> and <limits> are unnecessary includes in
some of the files?

Otherwise, it worked, and was much faster than std::sort(), so well done.
 :)

I'm not real sure about the interface, though: do they have to be different
named functions (integer_, float_, string_)?  Can't it be called
bucket_sort() and specialized for different types?  And reverse sorting,
isn't that done by passing reverse iterators?

Out of curiosity, I decided to tackle the problem and wrote an
implementation of counting sort, which as you may know, is one way of
implementing radix sort.  The code is not ready for a formal review but I
would really appreciate feedback on it, as it is written with a view to
inclusion in Boost.

https://github.com/jeremy-murphy/integer-sort

I'm particularly still uncertain about the interface (asking for forward
AND reverse iterators seems annoying, however legitimate).  Thanks, cheers.

Jeremy

On 14 July 2013 01:57, Sourav <souravatboost_at_[hidden]> wrote:

> >> From: Boost [mailto:boost-bounces_at_[hidden]] On Behalf Of
>
> > >> Steven Ross
> > >> Sent: Thursday, July 11, 2013 11:52 AM
> > >> To: boost_at_[hidden]
> > >> Subject: Re: [boost] radix sort
> > >>
> > >> Here's the boost library ready for review:
> > >>
> >
> https://github.com/boost-vault/Miscellaneous/blob/master/algorithm_sorting.zip
> > >>
> > >> And here's the sourceforge project (same code, but setup for running
> > >> outside of boost): http://sourceforge.net/projects/spreadsort/
> > >>
> >
>
> Steven,
>           Thanks for the response. I am just trying to pick up a project
> from the algorithms/data structure
> sections listed in the  SoC page.  I do not know whether there is a demand
> for the regular radix sort.
>
> Regards,
> -Sourav
>
> _______________________________________________
> Unsubscribe & other changes:
> http://lists.boost.org/mailman/listinfo.cgi/boost
>