Re: [boost] Re: shifted_ptr review

From: Ronald Garcia (garcia_at_[hidden])
Date: 2003-09-29 16:01:58

Philippe A. Bouchard wrote:

>My apologies for the documentation.  I didn't cover some areas like missing
>placement operators, etc. because the solutions are based on theoretical
>extensions: page mapping for heap allocations from which you can quickly
>detect the type of the pointee object (placement new used, global allocator,
>garbage collected object, etc.).  This would put an end to allocations
>failures (seg faults).
>
>I am going to rewrite the documentation, but the main idea was at least
>mentionned ;)
>
>  
>
I was planning to review this library, but I find myself quite 
discouraged by the comments regarding the documentation.  I feel that 
substantial documentation should exist for a library
before it is considered ready for review.  I recognize that there's a 
certain risk involved in heavily documenting a library prior to review: 
It may change a great deal as a result, or may not even be accepted.  
Nonetheless, a statement of rationale and library interface reference 
documentation has the potential to improve the quality of reviews and 
increase their number as well -- reviews based on documentation alone 
can be quite valuable.

My 0.02

ron