Comments on Andrew's mail

Hello,

I would like to comment the email of Andrew Woolf, sent to the DODS mailing 
list dealing with geomatics in general and OpenGIS and ISO more specifically.

I completely share the opinion of Andrew. I also consider that the real 
question is about data modelling and encoding.

For me an open issue is whether we should introduce a unique general model 
(see for example the GML 3.0 model), or to work  
on models interoperability (i.e. ontology/semantic stuff). As a matter of fact 
the OpenGIS already presents three  
different services characterised by diverse  models: the WFS, WCS and WMS.
As far as WCS, it's great in order to import gridded-based datasets into GIS; 
as a matter of fact, we are implementing a  
WCS interface for DODS in the THREDDS project.
The OpenGIS has been essentially driven by GIS Community for several years, 
only recently Atmospheric and Climatology  
Research Communities are active in the Consortium. Therefore, I think WCS is 
still in his infancy compared to WFS or WMS.
> From my point of view, there is still some work to do on web services and XML 
encoding for gridded-based datasets; for  
example, an interesting work under development is that of the NcML initiative 
or ESML. But here the issue of the model  
returns.

I'm sure that web services and grid services are playing and will play a 
decisive role in bringing together GIS and  
Atmospheric research Communities, and Science and Society in general. As a 
matter of fact, the OpenGIS Consortium is active  
since 1994 and around the 1996 issued a sort of Feature service on the top of 
CORBA, Microsoft DCOM, and SQL but its real  
popularity mainly arrived with the Web revolution and the OWS initiatives. 
The OpenGIS work is great, but I think that the collaboration with ISO TC211 
is extremely important to guarantee a  
world-wide consensus and the required openess and formalisation: essentially, 
what I consider important is to push  
interoperability. 

--Stefano

P.S. I attach the Andrew's mail for documentation.

----------------------------------------------------
Subject:
RE: discussion about DODS & MapServer
> From:
"Woolf, A (Andrew) " <address@hidden>
Date:
Tue, 24 Jun 2003 10:36:41 +0100

To:
address@hidden

I think this is a fantastic and timely discussion. There is a growing
convergence between the sorts of things we're wanting to do in the climate
sciences and what's going on in the GIS community. Parallels I see include
LAS / OGC Web Map Service and OPeNDAP / OGC Web Coverage Service. I'm also
convinced that the OGC Web Services strategy in general is a crucial
development. Web Services are becoming very widely accepted as a standard
architecture for distributed systems. They also happen to form the basis of
"Grid Services" (and there are a number of international "Grid" projects in
the earth sciences).


With regard to the georeferencing issue, the main problem is simply the
limited models developed by the GIS community so far for what they call
"Grid coverages" and gridded datasets.


For instance the "Grid Coverage" types defined in OGC Abstract Specification
Topic 6
(http://www.opengis.org/techno/abstract.htm) and the Geography Markup
Language
(http://www.opengis.org/techno/implementation.htm) provide only for
regularly spaced grids, although the OGC Grid Coverage Implementation
Specification allows for an affine relationship between indices and
coordinates.


The ISO Technical Committee 211 is developing a raft of standards (the 191xx
series) for geographic information. As with OGC, released standards to date
don't handle well gridded data. Their Tech Report ISO 19121 pointed to the
need to address this, and further work is underway (see work items 19124,
19129, 19130 at http://www.isotc211.org/scope.htm). In practise, OGC and
ISO/TC211 work very closely together and so OGC developments feed into ISO.
This is important because OGC right now has an initiative underway (OGC Web
Services 2, http://ip.opengis.org/ows2/) with one of the core themes being
the need to handle meteorological and hydrological model output.


Also of interest might be the European COST-719 initiative looking at the
use of GIS in climatology and meteorology
(http://www.knmi.nl/samenw/cost719/).


I think this is a very important area, and payoffs will be tremendous, but
there's a lot of work to do, and the more dialogue between the two
communities the better.


- Andrew