Roll Call

[Rob Atkinson]

Rob Atkinson,

My underlying interests are public access to public information, better 
informed science and policy etc. A functioning Spatial Data 
Infrastructure is required, and interoperable services and accessible 
tools. Have been exploring the richer end of "bind" - looking at the 
business, technical and theoretical issues behind WFS and processing 
chains and consequently done a lot of work in registry roles within SDI 
architectures.

Specific interests regarding clients are:

 I believe the plugin data issue is trivial at a protocol level, hard at 
a semantic level. Thus, a primary role of a common reference 
implementation (API if you like) is for interacting sets of components 
to realise an underlying information model. i.e. I agree with other 
sentiments about "mass-maps" providing impetus, but having a way too 
restricted view of the world.  What is our "view of the world" though?

More than plugging in data, however, is the concern about how modules 
can be plugged into a larger business application. Its all very well 
dropping in a map via a simple API, but how do you integrate the ability 
to discover and bind to a new data source? Or, in the more concrete 
case, be able to guide a user through interaction with a WFS that may 
have terabytes of data behind it?

We need to tease out the common "contract" between client behaviour and 
service semantics, and encapsulate this an API.  At the moment we do 
this partially - for the spatial navigation side of the contract, but 
not for any other "business oriented" interactions.

What I'd like to see is bindings between the common tools ("widgets" if 
you like) and UI frameworks. Portlets is one obvious target.  Then, it 
should be possible to have multiple code bases (including proprietary 
ones) but not be limited by the functionality of any one implementation. 
For example, why not be able to switch between a 2D view and a 3D view 
in a completely different technology?


Rob Atkinson
Social Change Online