RE: Semantics in WSDL-S ?!

 

-----Original Message-----
From: jeff@inf.ed.ac.uk
To: public-sws-ig@w3.org
Sent: 11/21/05 2:11 PM
Subject: RE: Semantics in WSDL-S ?!


Quoting "Shi, Xuan" <xshi@GEO.WVU.edu>:

> For example, WSDL interface:
>
> Function intersect (String Polygon1, String Polyon2): String Polygon3.

It looks to me like the description of "intersect" above does
give the inputs and outputs and their types.  The types all
happen to be String.  Or am I misunderstanding the notation?


Yes, the input types are all strings in which geospatial polygon feature
layers are specified. Open GIS Consortium (OGC) formulated an industrial
standard or agreement as Geographic Markup Language (GML) for the GIS
community. For the purpose of prototype demonstration, I used a simplified
format in my development which can be accessed at:
http://157.182.136.51/agswsprojs/geoWSviewer/WebForm1.aspx

This application may show us how spatial features are defined and used in
Web services. In a similiar example, I developed a geospatial Web service
that performs the following function supported by ESRI's ArcGIS Server:

select polygon features from Input Polygon feature layer 1 that are
contained in Polygon feature layer 2.

This application can be accessed at:
http://157.182.136.51/agswsprojs/geoWSviewer/WebForm2.aspx

In this example, if requesters change the order of input polyons, they will
get a different answer as they will select polygon features from Input
Polygon feature layer 2 that are contained in Polygon feature layer 1. So we
have to define the meaning of the service and functions more explicitly and
this is why semantics is so important in geospatial Web services.

In all such examples, like the function Intersect, String polygon1, polygon2
and polygon3 may have the same semantics when this function is used to
generate a new feature dataset. This is because all input and output
polygons follow the same specifications, i.e. they have the same coordinate
system and datum for survey measurement. In nature, they are all geospatial
polygons sharing the same domain ontolgy while they use different names in
certain functions. If we try to add semantic annotations onto these three
polygons, such annotation will be the same and then there is a redundancy in
such process. 

How can we differentiate such polygons with the same semantics in one
functional interface in WSDL since WSDL-S has to target at a certain object
inside WSDL document while such requirement is associated with two or more
elements?

Received on Monday, 21 November 2005 20:30:43 UTC