1 Spatial Data on the Web: Issues and Challenges
5
simplification of triangle meshes and specific out-of-core multiresolution models for
regularly distributed data and irregularly distributed data.
Progressive data transmission is presented in Chap. 4. First, the author illustrates
well-established implementations for progressive transmissions of data over the Internet, including raster images and triangular meshes. Then the problem of transmitting geographical map data in vector format is analyzed and an overview of the prototype systems developed and discussed in the literature is presented. Moreover, the
author presents a critical discussion of the main research and implementation challenges still associated with progressive vector data transmission for Web-mapping.
Finally, in Chap. 5, a specific new approach for the generation and display of
spatial information at a given resolution level is presented. It is based on a variable
scale data structure, the topological generalized area partitioning (GAP) structure
called tGAP. The purpose of this structure is to store the geometry only once, at the
most detailed resolution level, and to represent in addition the result of a generalization algorithm applied to the stored geometry. Then, the tGAP is used when data
are requested by a Web client for deriving on-the-fly the geometry at the requested
resolution level, based on the generalization preprocessing.
1.2 Integration of Spatial Data Sources
A geographical data set is an abstraction of the real world, according to a specific
point of view and purpose. From this consideration, it follows that different processes (e.g. social, ecological, economical), based on different motivations, at different times (e.g. every 5 years), and possibly using different devices, may produce and
may make available online different geographical data sets, concerning the same or
overlapping areas to users and applications. In a distributed environment, several application contexts may require the ability to use together and compare such distinct
data sets. The process of making distinct data sets usable in a homogeneous way by
a given application is called “data integration process”.
Integration processes should be able to cope with heterogeneity concerning how
source data sets are represented and how they can be accessed. Heterogeneity in data
representation arises when the same concept is represented differently in each data
set, thus generating some semantic conflict. Conflicts can be related to the meaning
assigned to concept names (same meaning but different concept names or different
meaning and same concept name), descriptive attributes used to describe a concept
(a road can be associated with an attribute length in a data set and with attributes
length and type in another), data types assigned to spatial attributes associated with
concepts (a road can be a polygon in a data set and a line in another), and units
used to represent geometric attributes (road lengths can be expressed in kilometers
in one data set and in meters in another). Some of the conflicts cited above, for
example those related to concept meaning and descriptive attributes, are not specific
to the spatial context, and can therefore arise whatever the application domain. On the
other hand, those related to geometric attributes and spatial concepts are typical of the
5
simplification of triangle meshes and specific out-of-core multiresolution models for
regularly distributed data and irregularly distributed data.
Progressive data transmission is presented in Chap. 4. First, the author illustrates
well-established implementations for progressive transmissions of data over the Internet, including raster images and triangular meshes. Then the problem of transmitting geographical map data in vector format is analyzed and an overview of the prototype systems developed and discussed in the literature is presented. Moreover, the
author presents a critical discussion of the main research and implementation challenges still associated with progressive vector data transmission for Web-mapping.
Finally, in Chap. 5, a specific new approach for the generation and display of
spatial information at a given resolution level is presented. It is based on a variable
scale data structure, the topological generalized area partitioning (GAP) structure
called tGAP. The purpose of this structure is to store the geometry only once, at the
most detailed resolution level, and to represent in addition the result of a generalization algorithm applied to the stored geometry. Then, the tGAP is used when data
are requested by a Web client for deriving on-the-fly the geometry at the requested
resolution level, based on the generalization preprocessing.
1.2 Integration of Spatial Data Sources
A geographical data set is an abstraction of the real world, according to a specific
point of view and purpose. From this consideration, it follows that different processes (e.g. social, ecological, economical), based on different motivations, at different times (e.g. every 5 years), and possibly using different devices, may produce and
may make available online different geographical data sets, concerning the same or
overlapping areas to users and applications. In a distributed environment, several application contexts may require the ability to use together and compare such distinct
data sets. The process of making distinct data sets usable in a homogeneous way by
a given application is called “data integration process”.
Integration processes should be able to cope with heterogeneity concerning how
source data sets are represented and how they can be accessed. Heterogeneity in data
representation arises when the same concept is represented differently in each data
set, thus generating some semantic conflict. Conflicts can be related to the meaning
assigned to concept names (same meaning but different concept names or different
meaning and same concept name), descriptive attributes used to describe a concept
(a road can be associated with an attribute length in a data set and with attributes
length and type in another), data types assigned to spatial attributes associated with
concepts (a road can be a polygon in a data set and a line in another), and units
used to represent geometric attributes (road lengths can be expressed in kilometers
in one data set and in meters in another). Some of the conflicts cited above, for
example those related to concept meaning and descriptive attributes, are not specific
to the spatial context, and can therefore arise whatever the application domain. On the
other hand, those related to geometric attributes and spatial concepts are typical of the