|
Boost : |
Subject: Re: [boost] sorting library proposal (Was: Review Wizard Status Report for June 2009)o
From: Steven Ross (spreadsort_at_[hidden])
Date: 2009-08-17 16:37:26
On Mon, Aug 17, 2009 at 12:48 PM, DE <satan66613_at_[hidden]> wrote:
> on 17.08.2009 at 22:45
> Steven Ross wrote :
> > I have completed reformatting and updating the documentation and adding a
> > performance calculation/proof. My proposed Boost.Algorithm.Sorting
> library
> > is ready for review.
> > The library is available at:
> >
> http://www.boostpro.com/vault/index.php?action=downloadfile&filename=algorithm_sorting.zip&directory=&
> if you ask me i say
> 1. doc navigation is odd and inconvenient
I followed the conventions in:
http://www.boost.org/doc/libs/1_34_1/more/writingdoc/index.html.
Do you have any specific suggestions for improvement?
>
> 2. integer constants
> MAX_SPLITS
> MAX_FINISHING_SPLITS
> MIN_SORT_SIZE
> LOG_MEAN_BIN_SIZE
> LOG_MIN_SPLIT_COUNT
> LOG_FINISHING_COUNT
> FLOAT_LOG_MEAN_BIN_SIZE
> FLOAT_LOG_MIN_SPLIT_COUNT
> FLOAT_LOG_FINISHING_COUNT
> must be members of a (nameless) enum
Why? I don't see that in:
http://www.boost.org/development/int_const_guidelines.html
They are details that are not intended to be modified by the average user,
and are in the detail scope for that reason.
People with a good understanding of both the algorithm and the processor(s)
they intend to run their software on may wish to modify them, which is the
only reason they are documented.
>
>
> 3. i think it's a valuable piece of software for specific domain
>
> Thanks.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk