-------------------------------------------------------
Ing. Simone Giannecchini
GeoSolutions S.A.S.
Owner - Software Engineer
Via Carignoni 51
55041  Camaiore (LU)
Italy

phone: +39 0584983027
fax:      +39 0584983027
mob:    +39 333 8128928


http://www.geo-solutions.it
http://simboss.blogspot.com/
http://www.linkedin.com/in/simonegiannecchini

-------------------------------------------------------



On Tue, Jun 9, 2009 at 12:27 PM, Andrea Aime<[email protected]> wrote:
> Simone Giannecchini ha scritto:
>> I have a few doubts about the update of ReferencedEnvelope. Are you
>> sure it is already capable of doing 3D semantic-wise?
>> It implementes BoundingBox which is purely 2D and that's why it is
>> used almost everywhere in the feature code. If we start to put 3d
>> envelopes in it, we would be clearly violating the BoundingBox
>> interface or, put it another way, what is going to be the meaning of x
>> and y with a 3D crs ( it might seem a naive question and probably is,
>> but I am trying to understand which assumptions we are making).
>
> ReferencedEnvelope extends jts envelope and implements
> org.opengis.geometry.Envelope and BoundingBox
>
> JTS Envelope and BoundingBox are 2 dimensional,
> org.opengis.geometry.Envelope is n-dimensional and we implement
> the getDimension() and getMimumum(ordinate)/getMaximum(ordinate)
> that do allow to grab dimensions above the second.
>

This is correct of course, but right now this works "semantically"
because a ReferencedEnvelope IS a BBox.
If we proceed this way, this would not be true anymore, therefore
something that until yesterday was 2D would become ND.
I understand this might look like a sterile discussion, but I think it
is a sensitive topic give the vastity of places where the
ReferencedEnvelope is used.

> I was thinking that extending the class to n dimensions produces
> no harm, whoever is using it as a 2d one just won't notice the
> extra dimensions.
>
> I'm still doubtful about methods that do check relationships
> between envelopes thought. What should we do when a nd and
> a 2d envelope are compared? Flatten the nd one or return a
> marker that the comparison does not make sense.
>

This again is a sensitive topic since we do not have methods for going
from a 3D crs which is not compound to a 2D + 1D crs hence we cannot
always split the envelope.


> The main reason to have a 3d envelope class is to be able
> and return the 3d envelope of 3d data sets as well.
> As an alternative I guess I could sublcass ReferencedEnvelope,
> but that would create even more confusion...


 It would probably make sense. Would it add a lot of load on you shoulders?

Simone.
>
> Thoughts?
> Cheers
> Andrea
>
> --
> Andrea Aime
> OpenGeo - http://opengeo.org
> Expert service straight from the developers.
>
> ------------------------------------------------------------------------------
> Crystal Reports - New Free Runtime and 30 Day Trial
> Check out the new simplified licensing option that enables unlimited
> royalty-free distribution of the report engine for externally facing
> server and web deployment.
> http://p.sf.net/sfu/businessobjects
> _______________________________________________
> Geotools-devel mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/geotools-devel
>

------------------------------------------------------------------------------
Crystal Reports - New Free Runtime and 30 Day Trial
Check out the new simplified licensing option that enables unlimited
royalty-free distribution of the report engine for externally facing 
server and web deployment.
http://p.sf.net/sfu/businessobjects
_______________________________________________
Geotools-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/geotools-devel

Reply via email to