Re: [boost] [GGL] Initial thoughts; how to register point types

Subject: Re: [boost] [GGL] Initial thoughts; how to register point types
From: Barend Gehrels (barend_at_[hidden])
Date: 2009-11-14 17:35:28

Hi Rutger,

>
>> The whole library is designed like this and I think it is good and
>> generic. You're now able to use all functions in template environments,
>> where you don't know if you're handling a point or a line. If you write
>> distance_pt_to_poly, you have to know the context...
>>     
>
> I guess I have been referring to this overloading, too. IIRC, a different 
> distance will be returned for points at the exact same location but in 
> different coordinate systems? For example, I guess I would prefer 
>
> great_circle_distance( a, b );
>
> over a strategy-based overload of distance() returning the same, as long as 
> the strategy is defined by a coordinate system (and not by a space or a 
> metric).
>   
This is indeed how it is implemented,  "distance" everywhere. You can 
call "distance" for two points in a lat-long coordinate system and get 
back this great circle distance. Calling it for cartesian and you get 
back pythagorean distance.
So I understand you prefer to have it *not* the same. Sorry, I didn't 
understand this point from your previous posting. It works everywhere 
like this, you can call area and you get the area (in meters) which 
makes more sense than the area in square degrees. If you *don't* want 
that behaviour, you can override it with a strategy... Or force points 
to be in a different coordinate system than they really are...

Regards, Barend