[Geotools-devel] AW: ISO geometries

[<Mar...@bu...>]

All,

many thanks for the intro, Rob. To introduce ourselves a bit more, =
Andreas and myself will work with you regarding the harmonization. =
Andreas is the lead developer of deegree and knows the technical stuff =
better than I do. These harmonization issues will surely be a lot of =
work but I hope that both projects will benefit from it. Ultimately, if =
this harmonization work is successfull, deegree would become something =
like GeoServer, a project using GeoTools and developing OpenGIS Web =
Services on top of it.

In the CITE project we will try to harmonize the data model (Feature and =
Geometry classes). To get this work started we should narrow the =
discussion down to the geometry and to narrow it even more down I would =
propose to start with the interfaces. deegree uses interfaces based on =
ISO 19107/OpenGIS Abstract Specification Topic 1 (see =
http://www.opengis.org/techno/abstract.htm).=20

If I understand right, GeoTools is using JTS for its Geometry =
definition, so the model is simple features based. The interfaces are in =
the package com.vividsolutions.jts.geom. If I say anything about GT2 =
that I misunderstood, please correct me. I looked at the javadoc that =
can be found at http://www.vividsolutions.com/jts/jtshome.htm. Is that =
the right place?

The ISO 19107 interfaces we are using are different to those interfaces =
in two regards. One is that the inheritance hierarchy is different, =
using some intermediary interfaces before we get down to implementable =
interfaces. This will be the hard part to resolve. The second difference =
is that 3D-geometries are allowed (at least on interface level), so =
additional interfaces would be needed for GT2. This should be fairly =
easy to resolve. ISO 19107 supports all kind of other things as splines =
and topology for example. deegree doesn't support this so far, but the =
design of the interfaces would allow to do this in the future.

These differences should be documented with the help of UML diagrams I =
think. I would therefore ask the GT2 team to provide a UML diagram of =
the geometry interfaces (and classes) and we will provide one for the =
deegree geometry. From this point on we can get into details.

The support of those more complex geometries has several advantages. The =
main one is that GML3 is based on these classes and therefore we need =
those if we want to be able to handle GML3, what surely is an important =
thing to do in the future.=20

Ian, Rob, James (virtually...) and myself had some (more detailed) =
discussion about these harmonization issues at the last OGC meeting in =
Thousand Oaks already, but I wanted to give a general introduction first =
to get everyone involved.

I would also like to propose to put these harmonized interfaces right =
away in the GeoAPI process. James: what do you think about this? Can we =
get GeoAPI started during the next weeks?


Let me know what you think about this. I am looking forward to this work =
with you guys.


cheers

Markus




-----Urspr=FCngliche Nachricht-----
Von: rh...@op... [mailto:rh...@op...]
Gesendet: Dienstag, 28. Januar 2003 07:38
An: geo...@li...
Cc: M=FCller, Markus
Betreff: ISO geometries


All,

As you may know, as part of the OGC CITE initiative, we are moving to
harmonize many of the internal GT2 architecture with that of deegree.=20
Markus is one of the deegree developers and he will follow this email up
with a brief description of the differences in geometry representation
between deegree and GT2.  Just wanted to give him an intro and an =
entree.

The involvement of the deegree team is an exciting thing; they have a
great open source project and lots of developer talent.  We hold out
great hope that thier minds and fingers will accelerate the entrenchment
of GT2 as the leading OpenGIS, open source toolkit.

Although not all GT2 developers are part of the CITE process, we are
conducting this harmonization effort out in the open, like everything
else with GeoTools, and we invite everyone to participate in it.  James
can give more info on CITE and its relation to GeoTools.

Rob