From: Fernando Cacciola (fernando_cacciola_at_[hidden])
Date: 2004-04-01 12:58:27
Aleksey Gurtovoy wrote:
> Tom,
>
>> There is an interesting class on the boost vault
>> called geometry2d. What is the status of this class.
>
> The classes in the header are basically what I would characterize as
> "everybody's geometry", in the sense that they codify a widespread,
> commonly shared understanding of how these basic geometry primitives
> should look like (some details notwithstanding).
>
> Within that framework, they are relatively well-thought and tested,
> and are generally usable.
>
> I do not intend to push them towards the review, though (even when
> I hypothetically have time for something like that), because IMO it's
> a waste to implement a whole host of geometrical algorithms/operations
> for some specific geometry classes, no matter how generic these are.
> The algorithms are still the same whether you are dealing with your
> own classes, structs coming from platform APIs, or 3-d party libraries
> -- or, for that matter, mixing these together. What is needed is not
> a yet another set of geometric data structures, but rather a truly
> generic, second-generation geometric *algorithms* library that
> will work on pretty much everything that fits in conceptually
> (rather than technically).
>
> SFINAE makes it possible, and we have something proprietary along
> these lines here at work. I'd by happy to elaborate more on specific
> questions if you are interested to know more.
>
Now, if I got you right, this is interesting...
I've been using my own generic geometric algorithms that operate on
arbitrary geometric objects via traits-based reflection.
But of course writing the traits is troublesome and actually unnecesary.
Are you using SFINAE for automatic reflexion?
Fernando Cacciola
SciSoft