[OSGeo-Standards] Re: OpenSearch Geo Spatial Relation

[Volker Mische]

Hi Carl,

I agree that all those operations should supported when you deal with a
relation between two geometries, e.g. in a filter.

But in OpenSearch Geo it's different. It's about intersecting one
geometry with a full data set with lots of geometries. So it's not a one
to one relation. Hence we should strip those out that doesn't make sense
in that context.

It has already been done in OpenSearch geo hence there are only 3
relations specified in the standard as opposed to the 9 ones specified
in the Rgenhofer paper. And I even take it a step further and say that
"disjoint" doesn't make much sense/adds any value.

Cheers,
  Volker

PS: Finally with CC to the OSGeo standards list (previously I was
sending from the wrong email address).

On 09/29/2011 06:55 PM, creed at opengeospatial.org wrote:
> To add to the discussion.
> 
> "Disjoint" is one of the spatial operators that are included in the SQL-MM
> and OGC Simple Features SQL standards. Other operator terms are "Equals",
> "Touches", "Within", "Contains", "Intersects", "Crosses", and "Overlaps".
> I may have missed one. These operators are consistent with and based on
> the Egenhofer "Binary Topological Relations Between Regions" (Egenhofer
> and Herring).
> 
> For consistency between standards, I would suggest Opensearch Geo provide
> the same spatial search operators.
> 
> Regards
> 
> Carl
> 
> 
> 
>> Hi Volker
>>
>> One good option is the OSGEO standards mailing list (also in light of the
>> MoU between OGC and OSGEO)
>> In fact we discuss all this a few months ago on that list where I posted
>> the OGC document
>> That draft document includes that optional spatial relation that was a
>> request coming from the CSW3 SWG
>>
>> This new feature is still open for discussion and this is whey you don't
>> see on the opensearch web pages
>> http://www.opensearch.org/Specifications/OpenSearch/Extensions/Geo/1.0/Draft_2
>>
>> Do you seen a problem in having the spatial relation has optional ?
>> In fact, if you don't want to support this queriable in a service you
>> don't include it in the URL template of the OpenSearch Description
>> document
>> (as stated in that document the default is “overlaps” )
>>
>> ciao
>>
>> pedro
>>
>> On Sep 27, 2011, at 8:53 PM, Volker Mische wrote:
>>
>>> Hi all,
>>>
>>> is there a better place to send mails about the OGC work on OpenSearch
>>> Geo? I've just used an old email to find out who to send it to.
>>>
>>> Anyway, this mail is about the "Spatial Relation" part of the
>>> specification. There is "overlaps, "contains" and "disjoint". I don't
>>> see the value of disjoint.
>>>
>>> This has several reasons:
>>> - I can't imagine a case where it is useful
>>> - it potentially returns a huge result set
>>> - it can always be modeled with a/several contains request(s)
>>>
>>> So for me it just an addition to the specification that makes sense at
>>> first sight, but not when you think about it more deeply. I'd prefer if
>>> "disjoint" would be removed from the spec, it doesn't add much value,
>>> but just burden on the implementers.
>>>
>>> Cheers,
>>>  Volker
>>>
>>
>>
> 
>