[georss] Design consideration in incorprating georss gml into environmental data schema

[Louis Sweeny]

I work with www.exchangenetwork.net basically a partnership for
environmental data web services. We have COI's who develop complex rich
sets of schema for the hard core data (like water quality)

Members have experimented with GML, but it turns out that 90% of our
services can get by with the georss gml subset so we are moving forward
with recommendations that our mish-mash of ad hoc tags for lat/long
points and line/areas be replaced with georss tags as people rev their
schema. 

I'm scratching my head about a few design issues (some of which the list
has already kicked around):

1) Is there any benefit to adoption of the "where" construct in complex
data schema? For example are there now or are there likely to be
services out there that would be able to scan a big file, with 7 levels
of hierarchy, find some "where" tags and take a guess as to what to do
with them? Or is the where construct envisioned just for rss feeds.

2) We are trying to make our data services (e.g. GetWaterQualityData)
play well with geo services like wfs, wms and geo-rss services, so that,
for example, one could find a monitoring station via an rss or wms
service and then use our data service to drill down to the rich data.
That is making us realize that we want to complement our data services
with wfs/wms and rss feeds, so that folk can get to the data however
they want.  Anything anybody can think of, re the WAY we incorprate gml
tags into these data services that would make this easier or harder?

Any thought appreciated. I'll share the example we come up with and
demos back with the list.

Thanks.