Re: [boost] [Review] GGL review starts today, November 5th

Subject: Re: [boost] [Review] GGL review starts today, November 5th
From: Thomas Klimpel (Thomas.Klimpel_at_[hidden])
Date: 2009-12-05 10:37:52

Barend Gehrels wrote:
> Thomas Klimpel wrote:
> > I actually collected many typos during my review, and corrected
>>  some in the corresponding source files, but I will have to send this
>>  in a separate mail. (And I probably should also elaborate more on
>> the points were the documentation could/should be improved.)
>
> They are welcome!

I decided that it's easiest to fix all found typos in the sources. Find attached the corresponding patch. So I won't elaborate more on the points were the documentation could/should be improved right now, but I may review the improved documentation when it becomes available.

> > I heard that GGL has its own mailing lists.
> > When GGL gets accepted as boost library, this mailing list should at
> > least be referenced in the "http://www.boost.org/community/groups.html"
> > section of the boost web page. It might also be nice if in this case, that list
> > would be open to discussions regarding the development of geometry
> > related boost libraries and questions about them in general.
>
> Certainly, the GGL mailing list is open to the public, and open for
> discussions on GGL and geometry.

So, now that GGL has been accepted, how about adding this link to "http://www.boost.org/community/groups.html"? Should I file a ticket for the boost-webpage to add that link?

> Thanks again, Barend

Congratulations to the fact that GGL has been formally accepted into Boost. I'm also happy to hear the you decided to go with the name Boost.Geometry. I'm also happy that you started discussing the handling of the extensions.

As you may guess, what I've done now is to cleanup/erase my notes that I took during the GGL review. One last possibly interesting observation that wasn't included in my review is that way the specializations for algorithms working on multi-geometry are handled can lead to quadratic runtime complexity behavior in some cases, for example when determining the minimum distance of two multi-point geometries. But to be honest, I don't think that this is really important, and if it were really important, it would still be possible to override the specialization of the algorithm for the cases were it is important.

Regards,
Thomas