[OSGeo-Standards] OWS Context and others OGC interactions

Lorenzo Becchi lorenzo at ominiverdi.com
Tue Oct 23 18:07:52 EDT 2007


I guess I have to apologize again for my poor English and probably for 
my poor OGC knowledge.
anyway...

Jody Garnett wrote:
> Hi Lorenzo; just want to make sure I understand the goal here ... if I 
> understand the goal here is how to add more stuff (specifically 
> support for additional protocols) to these documents?
yes, thanks for translating.

>
> One thing I liked about a lot of the OGC specs is their respect for 
> "vendor specific extentions"; the idea is to create a schema with a 
> well known place for people to dump in additional content; after a 
> while (if the need is common enough) you can promote the solution to a 
> first class element in your schema.

ok Jody, I possibly see the point. Tom example helped too.

my ideas is basically this:
To me OWS Context could be the Open Format that anybody is looking for 
to define map structures.  I mean: why should we deal with Mapfile, 
GeoServer configuration, OpenLayers configuration, uDig format (I guess 
you have one), QGIS format, gvSIG format, and all proprietary vendors 
formats when we can probably do everything with an OWS Context?

If we accept this point of view even OGC AbstractResourceTypes should be 
Extensions.

what happens when you export your map with local file-system links in?
I'm sure there are guys smarter then me here to answer, and we can act 
in many ways.
I know some were planning to create a simple zipped format with 
geo-metadata (or just metadata) inside to move such a kind of resource.  
It would be really cool to have GeoServer or Mapserver creating a 
WebService environment automatically from such a zip file. And the same 
for every Desktop GIS.

With OpenLayers naturally it is impossible to unzip or create a service 
in Javascript but we can have a "light" or "web" version of the Context.
Then your Desktop application should be capable to create both types of 
Context starting from the same map view. One for the server, one for the 
Web client.

To avoid confusion or "Fred's GML profile for rollerblading trails", all 
types of extension should pass through an incubation process as OSGeo is 
doing for its software.
This will allow new extensions to be tested an publicly available, then 
possibly implemented by all softwares.

am I totally out of Context?
;-)

ciao
Lorenzo





More information about the Standards mailing list