Well, yes, you're right that this isn't the answer I'd hoped for. But
the good news is I need look no further for an answer. I guess for the
time being I'll see about using a separate instance at 1.6 using the
community schemas as documented, and see how that goes.
As far as voices and priorities, I will say that I've been surprised
there isn't more demand for at least the simple case of xy pairs
representing point locations in a non-spatially-enabled database table
... surprised that it isn't a mainstream requirement by now to do what
the "locationsxy" code does. We (state government in New Jersey-USA)
are getting increasing demand for lightweight, simple mashup apps that
combine some particular business data set with a standard pre-rendered
base map. And in many of these cases, the canonical source for the
locations is some table or other in a database that is not managed as
spatial data, but does include location information. Traffic cameras,
park-and-ride lots, geodetic monuments, water wells, etc. We
conceptualized a connector that would respond to a WFS request by
retrieving records (possibly within a BBOX) from a table and delivering
them in GeoJSON to an Open Layers client app. We see it opening up a
huge variety of data t
o being displayed on a map, which will be the first foray into GIS for
many of these agencies.
Unfortunately, I have no resources to offer to help make it happen in
Geoserver. It's a tight time for us. So I will see about making it
work with what's out there.
Thanks for the quick reply.
-Andy
----- Original Message -----
From: Rob Atkinson <robatkinson101@anonymised.com>
Date: Wednesday, September 30, 2009 6:32 pm
Subject: Re: [Geoserver-devel] geometryless
Hi,
dont have the answer you are looking for I expect..
the _plan_ is to replace geometryless by the app-schema functionality.
This can done at the app-schema part - , but is held back by the
broader GML 3 support - we dont have an object model we can map into,
there is a point solution mapping 2D geometry objects into the
bindings. I've been trying to get this relaxed to allow a complete
solution for a while - maybe your voice can help with the priorities.
I cant fix it myself because there is auto-generated code in the loop,
and the autogeneration process is not yet repeatable.Making it work in 1.7 is feasible, but not as important to me as 2.0
support. You would need to port app-schemas sql-datastore as well, or
refactor the code to remove the dependency.Sorry if this is not good news, and I feel for you..
Rob