[udig-devel] Topologies Issues

[Vitali Diatchkov]

HellO!

Yes. I devoted lot of time to investigating Topology issues. In general, not
only in  fairway of GeoTools or UDIG.  Well, my MSc thesis is related to
these issues, something like "Spatial vector data processing with topology
consistency ..", that is not finished yet (long long process)..

Here is a key issue - a data model. Most of things in GIS area are based on
Simple Feature Model while it is really simple to deal with, process,
persist, request from, etc. For operations taking into account relationships
between features such as  spatial relations, topological predicates, Simple
Feature Model is not a best choice. Topology Data Model is  a solution. SFM
can be converted into TDM and back. While most of the geographical data
stores are based on Simple Feature Model (because of reasons above), there
is workaround trick: at any point we need a TDM, we can create it from SFM
(not an easy issue also, btw). 

During processing of the data with taking spatial relations into account,
like features editing   with topology preserving, what possible design I
see:
1) We have a Simple Feature Model.
2) Start edit feature with topology consistency. Here always an Area of
Interest can be defined (around feature, for example) where TDM is built
from SFM ob the fly.
3) Real editing tools deal with TDM (with nodes, edges, faces, etc.). On
this stage it is 200%  more easily to check various topological constraints
during editing, etc then if to work against SFM, also from performance point
of view. There are other advantages also.
4) When the editing is done, TDM is converted into SFM - means all
decomposed features are gathered back from TDM entities like node, edge,
face, etc (depending on geometry type of initial SFM).

What are weak places:
1) There is no good and standardized implementation of TDM on Java. There
are attempts in different projects but there is no good, based on open
standard and full enough at the current moment (like JTS geometry model).
2) The step of decomposing SFM into TDM is not that trivial, there various
optimizations are possible depending on SFM's geometry type.
3) As a consequence there are no tools and algorithms performing over TDM.


What I am telling concerns open source world, proprietary implementations
exist of course (ESRI, Oracle, etc, but they are proprietary).


Also. It is possible to perform operations where spatial relationships are
taken into account without TDM, having and dealing only with SFM. I can
count lot of disadvantages of this approach. Actually there is a good
document by ESRI about topology data model importance and how it makes a
life easier.

Hope it makes sense - if you wanna do the whole path - there are lot of
intermediate things to be done. The first one is an open source TDM if you
are intending to work with open source GIS.

Best wishes,
 Vitali Diatchkov.

> -----Original Message-----
> From: udig-devel-bounces at lists.refractions.net [mailto:udig-devel-
> bounces at lists.refractions.net] On Behalf Of Victor Mauricio Pazos
> Sent: Friday, September 08, 2006 12:56 PM
> To: User-friendly Desktop Internet GIS
> Subject: [udig-devel] Topologies Issues
> 
> Hi, we are thinking to develop some issues to hand features with
> topological
> relations in uDig.
> 
> issue1: Editing topological features.
> 
> issue2: Process to establish spatial relationships.
> 
> Someone is thinking or developing theses issues or related issues?
> 
> Thanks in advance.
> 
> --
> Mauricio Pazos
> Director
> 
> Axios Engineering
> www.axios.es
> e-mail: maurcio.pazos at axios.es
> tel-:+34 94 441 63 84
> fax: +34-94 441 64 90
> _______________________________________________
> User-friendly Desktop Internet GIS (uDig)
> http://udig.refractions.net
> http://lists.refractions.net/mailman/listinfo/udig-devel