[GRASS-user] survey: use of different geometry types in same vector map

[Helmut Kudrnovsky]

Markus Metz-3 wrote
> On Wed, Mar 5, 2014 at 12:08 PM, Moritz Lennert
> <

> mlennert at .worldonline

> > wrote:
>> Hello,
>>
>> Recent discussions here with colleagues about GRASS' vector format and
>> the
>> teaching of vector handling in GIS have brought up a question about the
>> fact
>> that GRASS (contrary to some other well-known vector formats) allows a
>> mix
>> of geometry types (points, lines, polygons) in the same map, something
>> which
>> some GIS'ers consider quite unorthodox.
> 
> Nothing to add to Ben's reply from a user perspective.
> 
> From a conceptual perspective, it depends on the definition of "vector
> map". Considering the number of vector formats (76 supported by OGR),
> any definition must be independent of a specific vector format.
> 
> In OGR, a vector is defined as a data source which contains any number
> of layers which in turn contain vector geometry objects. A vector
> geometry object in turn can be a collection of vector geometry objects
> of possibly different types. This property of a geometry object and
> geometry collection conforms to the OGC simple feature specification,
> a software and vector format independent (!) standard. That means that
> even a single geometry object in a single layer in a single data
> source can contain multiple feature types.
> 
> In my experience, a common task (probably the most common task) is to
> convert GIS data of any format to the format of your choice. Thus I
> would be surprised if regular GIS users are not familiar with the
> diversity you find out there in the wild.
> 
> Markus M
> 
>>
>> In order to enrich the reflection on that, I would like to ask GRASS
>> users
>> for use cases where this mix has been useful. Do you use such mixed
>> geometry
>> maps ? Which specific use cases do you use them for ?
>>
>> In order not to flood the mailing list, you can also send me your
>> response
>> by private mail. I'll report back to the mailing list with the results.
>>
>> Moritz
>> _______________________________________________
>> grass-user mailing list
>> 

> grass-user at .osgeo

>> http://lists.osgeo.org/mailman/listinfo/grass-user
> _______________________________________________
> grass-user mailing list

> grass-user at .osgeo

> http://lists.osgeo.org/mailman/listinfo/grass-user

I concur with all Ben and Markus have said.

some personal experiences as vegetation scientist who has done a lot of
habitat mapping for national administrations and agencies: 

habitat entities can be points (e.g. old trees, ...), lines (e.g. small
running waters, ...) or polygones (hay meadows, ...); in all the cases the
GIS representations of points, lines and polygones were kept in different
GIS datasets (shapefiles), although all these entities are often related to
each other and share often the same database schema. 

it was always a pain to keep three different GIS dataset updated or to amend
the database three times...

this is caused by historical technical reason, but for habitat mapping IMHO
this should be avoided in future because technical constraints have be
changed ...





-----
best regards
Helmut
--
View this message in context: http://osgeo-org.1560.x6.nabble.com/survey-use-of-different-geometry-types-in-same-vector-map-tp5107649p5107779.html
Sent from the Grass - Users mailing list archive at Nabble.com.