[postgis-devel] Polyhedral Surface / Polyhedron /TINS

[Vincent Mora]

On 10/07/2014 09:57, Olivier Courtin wrote:
> On Jul 10, 2014, at 1:45 AM, Paragon Corporation wrote:
>
> Regina,
>
>> I think this is a question for you or Vincent Mora.
>>
>> I recall you saying there is a bit in the PolyhedralSurface of PostGIS that
>> denotes if it represents a volume (Polyhedron).  Is that true or was I
>> mistaken.
> Yes there's a SOLID bit, meaning that the surface is closed,
> and so enclose a volume.
>
>> Also some other questions what are the rules of validity for Polyhedral
>> surface and TIN.
>>
>> I've always assumed that TINs are really a special class of Polyhedral
>> Surface (one composed of all triangles instead of varying size polygons) and
>> so the rules that apply to one should be similar to the other.
>>
>> I'm guessing
>>
>> 1) For both of course the polygon patches and triangles can't overlap
> Indeed,
>
>> The others not so sure
>> 2) They must share an edge with at lease one other (no islands) -- not sure
>> if this is true or not?
> There's a no reason for such a thing (must sharing),
> A TIN with a single Triangle inside is valid for instance.
>
> But indeed it's the way we check, for now, if the surface is closed or not.
>
>> 3) Are they allowed to share the same edge with more than one other polygons
>> and triangles. I wasn't quite clear from CGAL.  Seemed like it was legal
>> just a rare occurrence and more representative of a complex polyhedral
>> surface.
> Humm that's a good question, because if yes the previous closed test
> will not fit in this case...
>
> Vincent do you have any input on this one related to (SF)CGAL ?
> I will look further in OGC/ISO spec,
>
> O.
Not legal according to SFS rev 1.2.1 
(http://portal.opengeospatial.org/files/?artifact_id=25355), page 29, 
first line 1 :

  "...such LineString shall be part of the boundary of at most 2 Polygon 
patches"

V.