Subject: [boost] GGL
From: Barend Gehrels (barend_at_[hidden])
Date: 2009-09-18 16:57:12
Sorry for my late reply, I was off. I'll react now on the GGL part and
therefore I changed the subject.
Phil Endecott wrote:
> Barend has presented a few previews of his library and I have looked
> at these, though
> not in much detail. My feeling was that Barend had a long way to go
> before he would
> have something useful, and that even then he may never produce
> something that I would
> actually be able to use.
Other people are using it already, geometry is a large subject, it is
hard to serve everyone. In your case, I'm sure you could use it.
> For example, he has decided to adopt the conventions of the
> OGC; this is not entirely unreasonable, but it means that semantics of
> things like
> whether border points are inside or outside a polygon are non-negotiable.
The library follows OGC conventions loosely, not strictly, and
implements more than OGC so the example you give will surely be
implemented. Actually there was a question on this on the GGL-list this
> Barend has also recently reminded us of the spatial indexes GSOC
> project and says that it is now
> included in his library, despite it being the worse code I have ever
> seen presented to
Please tone down a bit your critics. The student dropped the project
right after presentation because of these kind of statements. I
contacted him this summer and he was prepared to work on it again,
though busy. Besides this, it is revised and will be revised more. And
finally, it is plugged in, nothing depends on it (yet), you can use
another spatial index or please provide your own.
> Despite all that, Barend's recent comments suggest that he has made
> progress and that he may have something useful to show us in a few weeks.
There should be something again in October. Note also that it is not *my
*library but a group's library, we're working with at least three (being
Bruno, Mateusz and myself) on it, and more people are using it, thinking
about it and making suggestions.