Re: [boost] [gsoc] boost.heap update

Subject: Re: [boost] [gsoc] boost.heap update
From: Roland Bock (rbock_at_[hidden])
Date: 2010-07-29 04:24:18

Tim Blechmann wrote:
> hi all,
> 
> i've uploaded a new snapshot of boost.heap. it includes some of the proposed 
> changes and i have also added a benchmark sections.
> during the final time for the gsoc, i'd like to get the library ready for a 
> review ...
> 
> therefore, i uploaded a new snapshot [1] and an updated documentation [2].
> 
> feedback welcome ...

Hi,

some feedback on the documentation:

* Motivation: Many boost libraries offer a "Motivation" section to say
what real world issues are addressed by the library. Currently, I am
using std::priority_queue and it suits my needs. When would they be
insufficient?

* Navigation: Some pages are missing navigation elements, e.g.
http://tim.klingt.org/boost_heap/heap/concepts.html should have a
"top"-link.

* Warnings: For instance: Mutability: "Incorrect use of increase or
decrease  may corrupt the priority queue data structure!" It would be
good to have short examples of correct and incorrect usage.

* Data structure comparisons: The information seems rather incomplete.
For instance, "In contrast to d-ary heaps, binomial heaps can also be
merged in O(log n)." OK, now I know, that binomial heaps can be merged
in O(log n), and d-ary can't. But I don't know the complexity for
merging d-ary nor any other heap.
Maybe it would be easier to use one or more tables to compare the
features/complexities?

Regards,

Roland