Database Server Models for WMS and WFS Geographic Web Services - Geographical Information Systems: Trends and Technologies - page 152

It is of prime importance to integrate the use of WFS operations. In the
current state of OWS, a requirement based on the visualization of a huge
volume of data leads to use a WMS service. This service is better fi tted and
widely suffi cient. Furthermore, if the requirement is to access to specifi c
data, to modify on the fl y the representation of specifi c data or to update
them then WFS operations are mandatory. Nevertheless, this service must be
used with precautions since the involved volume of data may be important
and therefore the response time may be signifi cant.
The goal of our proposal is to improve OWS such as WMS and WFS,
mainly the GetFeatureInfo and GetFeature operations. The aims are to
provide a better relevance of alphanumeric data to improve spatial analysis
and to provide a data model schema to handle such information. The second
proposition we presented in this section deals with rules to determine a
conceptual schema but does not consider instantiations of such models.
This is the matter we want to tackle.
Data Conceptual Schema
A selection of a (geo)graphical object (e.g., city, restaurant) represented on
a map, is translated into a query that will retrieve information about this
object. This knowledge (schema and data) that can be considered as static,
is provided by default to all users. It does not correspond in any way to
users' point of view. We, therefore, suggest adapting the database schema
associated with a query result, by withdrawals of irrelevant attributes and/
or additions of relevant attributes. We propose a mechanism of refi nement
for a query. It is a way to improve the search performance of relevant
attributes. It can be based, for example, on a process that adds to users'
queries new attributes related to those initially present in these queries.
Provided information will be richer and more “dynamic” than with a
simple use of a WFS call. Indeed, some additional attributes are the result
of methods (here functions) applied, among other things, on the attributes
associated with a spatial object. These methods are completely transparent
for an end-user. Provided information will always be, by the end, presented
as a set of attributes.
Obviously, this new database schema cannot be obtained from current
distant database schemas. We are therefore interested in determining a more
sophisticated database schema that integrates semantic links, attributes and
methods. We stand here at the conceptual level and introduce a semantic
extension that will allow the defi nition in a database schema, of relevant
attributes and methods. This requirement is related to the impossibility
of defi ning external schemas considering the heterogeneity of potential
clients. Figure 5 illustrates our approach that consists in enriching the data
schema.