[postgis-devel] [mtnclimb at gmail.com: Re: Should EMPTY be spatially equal to self ?]

[Sandro Santilli]

I think this was meant to go to postgis-devel as well...

----- Forwarded message from Martin Davis <mtnclimb at gmail.com> -----

Date: Tue, 17 Jan 2012 10:24:59 -0800
From: Martin Davis <mtnclimb at gmail.com>
To: Sandro Santilli <strk at keybit.net>, mateusz at loskoty.net
Subject: Re: [postgis-devel] Should EMPTY be spatially equal to self ?
In-Reply-To: <20120117094719.GS28071 at gnash>

strk, mateusz:

a couple of further comments:

- The T*F.F*F.FF* matrix was introduced in JTS as a correction to the one
given in the OGC 06-103r4 Simple Features Common Architecture document.  It
gives TFF.FTF.FFT, which must be wrong because it doesn't allow coincident
points to be equal.

- I wouldn't put too much faith in other spatial database specs, either.
 For one thing, even the standards docs can be wrong (as above).  For
another, I know that in the SQL Server case they used the JTS test suite to
validate their code!  (Although of course I'm sure if they thought
something was wrong they would have changed it)

One interesting thing I found while looking into how JTS handles this issue
is that there actually aren't very many tests for the named predicates in
the JTS Test Suite!  I was a bit shocked to find this, since I had always
thought this was tested extensively.  It does test the relate() method very
extensively, but for some reason the named predicates didn't get the same
attention.  I'm going to fix this, by adding a lot of tests for the names
predicates.  Also, there is no or little testing for EMPTY geometries -
something else I'll try and add.  This doesn't provide any insight into
what the semantics should be, but at least it allows changes to be tested
to see what their impact would be.

Martin

On Tue, Jan 17, 2012 at 1:47 AM, Sandro Santilli <strk at keybit.net> wrote:

>
> >
> > Strk -- where are you getting these definitions from?
>
> The EQUALS definition was reported by Martin Davis above.


> > I think the best way to resolve this is ot just do what the other spatial
> > databases do assuming they all do the same.
> > Just need some people to test SQL Server and Oracle.
>
> I tend to disagree about that but results I'm happy to look at.
>
> Disagreement is due to proprietary software often seen doing
> silly things due to lack of peer-review.
>
>

----- End forwarded message -----

-- 

  ()   Free GIS & Flash consultant/developer
  /\   http://strk.keybit.net/services.html