Geometry Boost List: Re: [geometry] concave hull

Subject: Re: [geometry] concave hull
From: Barend Gehrels (barend_at_[hidden])
Date: 2016-06-01 17:05:07

Hi Jeremy, Adam,

Op 1-6-2016 om 22:06 schreef Adam Wulkiewicz:
> Hi Jeremy,
>
> Jeremy Murphy wrote:
>> Hi there,
>>
>> a year or so ago I brought up (I can't remember exactly with who or 
>> where) the possibility of contributing a concave hull algorithm and 
>> the good news is that I'm close to convincing my work to open-source 
>> an implementation I wrote.
>>
>> To justify additional effort in preparing the code for release, they 
>> (senior leads) want to confirm that other people actually want it. I 
>> think that's hard for other people to give without seeing the actual 
>> code, but what I'm hoping for from the maintainers is just a 
>> non-binding "yes" that, in principle, you would be interested in an 
>> implementation of the concave hull algorithm described in the 
>> following paper:
>>
>> http://www.iis.sinica.edu.tw/page/jise/2012/201205_10.pdf
>>
>> There would be no obligation for you to accept or take custody of the 
>> code (obviously), they just want to know that we're not wasting our 
>> time preparing it. I originally wrote it with contribution in mind, 
>> so I really just need to do some polishing, make it respectable, etc.
>
> I personally think that every contribution is welcome. There is no 
> concave_hull() algorithm in Boost.Geometry and I think it would be a 
> valuable addition. So it's a "yes" from me.
> Barend, what do you think?

Yes, sure - thanks for the reminder, I indeed wanted to react on this. I 
also think that it is a very useful addition to our library. Actually in 
the past it has been asked for.

> Are there any users which would find this addition welcome?
>
> You probably already know this but just in case. We use certain 
> conventions at Boost.Geometry. Every algorithm is divided into two 
> parts, geometry-specific algorithm part dispatched by geometry tag and 
> coordinate-system-specific strategies. In the case of convex_hull() 
> the algorithm simply calls strategy which does everything. So I 
> suppose it could be the same in this case.
>
> As a part of your implementation have you used convex hull algorithm 
> implemented in Boost.Geometry or have you implemented your own?

Regards, Barend