Groundwork: Basic software installations
There have already been successful implementations of GeoSciML Web Feature services, using a few different approaches. Specifically, I'm aware of some work by Alistair Ritchie which uses customized configuration files for Deegree (a walkthrough is here), and also older work by Bruce Johnson using a combination of ArcIMS and Cocoon (another walkthrough here)
On top of that, I've heard that there has been a lot of work put into developing a GeoSciML WFS implementation in Geoserver. From playing around with it a bit, Geoserver is a nice package, and depending on how far along this development is (I don't speak any Java... yet), this would be very alluring.
With that in mind, the first thing for me to do is to install the required software to be able to experiment with these software packages. Below is a table of the software I installed, and links for them.
|Software||URL||What does it do?|
|Apache||http://httpd.apache.org/||Apache is an HTTP server. You can use it to host webpages.|
|Tomcat||http://tomcat.apache.org/||Tomcat is a servlet engine. It runs webapps like Deegree or Geoserver|
|Deegree||http://www.deegree.org/||Installed only the WFS package. This is one example of a WFS Server|
|GeoServer||http://geoserver.org/display/GEOS/Welcome||GeoServer is a combination WMS/WFS/WCS Server|
|PostgreSQL||http://www.postgresql.org/||PostgreSQL is a database management system. Note: I installed version 8.3.0|
|PostGIS||http://postgis.refractions.net/||PostGIS makes PostgreSQL handle spatial features. Note: I installed version 1.3.2|
In the end, I'm trying to build a workflow that looks something like this:
1. Generate my spatial data in my in-house production database. Here at the Arizona Geological Survey, we do this using ArcGIS software, and we do it in an SQL Server database.
2. Get the data from my proprietary, expensive, ESRI Geodatabase into the free, happy, open-source PostgreSQL + PostGIS environment. I want to do this using my expensive ESRI tools as much as possible. This involves setting up ArcSDE to work with PostgreSQL, and this is why I had to install out-dated versions of PostgreSQL and PostGIS. Soon I'll put together a walkthrough for making ArcSDE happily play with PostgreSQL, as it is not perfectly straightforward.
3. Use some open-source spatial data server (Deegree or GeoServer), and customized configurations to create a WFS that delivers my data in GeoSciML format. Someday I would like to see GeoSciML spit out of an ESRI product (ArcGIS Server now as opposed to ArcIMS), but that seems more trouble than it may be worth right now, and I'd also like to end up with something that someone could use even if they don't have access to ESRI products.