Subject: Re: [boost] Formal Review: Boost.Polygon starts today August 24, 2009
From: Thomas Klimpel (Thomas.Klimpel_at_[hidden])
Date: 2009-09-01 15:03:22
Simonson, Lucanus J wrote:
> Are you suggesting that I separate the declarations of functions
> from their defintions and place the defintions in the detail directory?
No. What I'm suggesting is that you differentiate between algorithms that can work on a concept and the interface of the concept. If I would implement an algorithm "max_axisparallel_ellipse" that would work on a polygon set, this wouldn't make it part of the interface of polygon set. But why should all algorithms implemented by you that work on a polygon set be part of the interface of polygon set, but not the algorithms implemented by me?
So what I suggest is that you separate the interface of a concept from the algorithms that work on the concept. Not even the algorithm dispatch code should be part of the interface of the concept.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk