[postgis-devel] Buffer style naming

[Paragon Corporation]

> I can do geodetic without disk changes by making it a fully parallel type,
at the cost of losing some code sharing in liblwgeom. It may well be that
this is the best thing, and then merge the types onto the same liblwgeom for
2.0. Although if we avoid the > disk change forever, ... there are many
things we won't be able to do.

Separation sounds better to me as a start anyway.  That way we can see how
it behaves in isolation.  But for 2.0 I agree we should rethink our
structure especially if we want to introduce all the point cloud stuff and
extended 3D support afterward.

> I am going to try and do my work on a spike and merge trunk into my spike
regularly so that I can merge back into trunk when I have something that
"does stuff". Until my hands are really dirty, the best course of action is
not going to be clear to me.
Great.  I'll be happy to test out stuff and provide moral support, though
I'm probably too stupid to help much more than that :(.

I forgot what was the reason we decided we can't use any of the indexing
support in pgSphere.  I was going to take a look at that since I have never
tried doing so before.  I mean seeing what's there and getting a sense of
what its capable of.
I see they just released a new version.
http://pgfoundry.org/projects/pgsphere/


>> Though I think you are dreaming if you expect the full blown geodetic 
>> and stress tested by December.  I look forward to be proved wrong though.

> It's only going to have 4 functions :) So hopefully I will prove you
wrong.
You always have before. I trust you will. :)

Thanks,
Regina