Re: [boost] [gsoc17]The implement of Programming competency test From Ruoyun Jing

Subject: Re: [boost] [gsoc17]The implement of Programming competency test From Ruoyun Jing
From: Ruoyun Jing (jingry0321_at_[hidden])
Date: 2017-03-20 12:51:22

Hi Vissarion,

2017-03-20 20:29 GMT+08:00 Vissarion Fisikopoulos <fisikop_at_[hidden]>:

> Hi Ruoyun,
>
> > - The main work I should do during GSoC about project 1 is just to
> improve
> > the algorithms of compare two distances of two point pairs or try to
> > consider different situations which used compare distance predicates to
> > avoid calculation?
>
> the project is about algorithms that does not compute the actual distance
> but an approximation (as rough as possible) that suffices to accurately
> compare the lengths of two geodesic segments on the ellipsoid
>
>
Thanks for your answer, I have the correct point of these project and will
continue to write my proposal.

> > - Do "In some algorithms there is the need to compare two distances of
> two
> > point pairs" this means compare two points distances on Earth? Or on
> every
> > ellipsoid model?
>
> the later (see answer above)
>
> > - Do I need to find the approaches of filter by myself? Could you please
> > help me get more information about these? Like the implement detail about
> > the predicate compare_distance function, mathematical calculation papers
> > about compare distance on ellipsoid?
>
> Well we give one approach on the website (there is another proposal of
> another
> candidate on the list). My opinion is that if you understand these
> approaches
> (pos and cons, differences etc) that will help you towards a solid
> application.
> Having another idea for an approach will definitely strengthen you
> application.
>

Thanks for your help. I can get the point of those approaches and will try
to have new ideas. :)

> > - In my view, just like compute the shortest distance between two given
> > points on Earth, the background approach can be used. And I used
> cartisian
> > coordinate to avoid many trigonometric function calculation, but I can
> not
> > find more improve with my lacking knowledge, could you give me some
> aspects
> > to get more approaches?
>
> Please see answer above
>
> At this point I think it is better to jump to boost list because our
> discussion
> could interest more people. I will not copy anything of our private
> conversation
> there but feel free to do the jump and I will continue answering there.
>
> Btw, any news from you test and comparison with Boost.Geometry.
>
> I reply you and jump to boost with your advice. Hope this will help more
people. :)

Vissarion
>
> ps. you can send questions whenever you want, I will answer them a.s.a.p.
> my zone is UTC+2
>
> Thanks for your answer.

Best regards.

Ruoyun
-- 
Northwest University of China
Software Engineering
jingry0321_at_[hidden] <jingry0321_at_[hidden]>