86
Marian de Vries and Peter van Oosterom
and edges (the GAP-edge forest). Faces and edges are assigned an importance range
(LOD range) for which they are valid. This set of tree-like structures is used when
the data is requested by a Web client to derive the geometry on-the-fly at the right
LOD, based on the importance values that were assigned during the generalization
process.
In the first part of this chapter we give a short overview of the tGAP structure. In
Sect. 5.2 the basic principles are described, followed by an explanation in Sect. 5.3
of how the generalization process builds a tGAP data set in a succession of steps.
In the second part of the chapter we explore how the tGAP data structures can be
used in a Web service/client environment. The focus is on two aspects: how the
tGAP structure can support progressive transfer of vector data from Web service to
client, and how adaptive zooming can be realized, preferably in small steps (‘smooth’
zooming). The relevant standards and protocols for vector data Web services are
discussed in Sect. 5.4: web feature service (WFS) and geography markup language
(GML). Section 5.5 explains how the tGAP structure can be used for progressive data
transfer and smooth zooming and proposes some necessary extensions to the current
standards (WFS and GML) in order to support vario-scale geo-information. Finally,
Sect. 5.6 concludes this chapter with a summary of the most important findings and
suggestions for further research.
5.1.1 Related Work
Vario-scale
Data structures supporting variable scale data sets are still very rare. There are a
number of data structures available for multiscale databases based on multiple representations (MRDBs) fixed number of scale (or resolution) intervals [7, 9, 25]. These
multiple representation data structures try to explicitly relate the objects at the different scale levels, in order to offer consistency during the use of the data. Most data
structures are intended to be used during the pan and zoom (in and out) operations
of a user, and in that sense multiscale data structures are already a serious improvement for interactive use as they do speed-up interaction and give a reasonable representation for a given LOD (scale). Drawbacks of the multiple representation data
structures are that they store redundant data (same coordinates, originating from the
same source) and that they support only a limited number of scale intervals.
Progressive Transfer
Another drawback of the multiple representation data structures is that they are not
suitable for progressive data transfer, as each scale interval requires its own (independent) graphic representation to be transferred. In a Web service/client context
progressive transfer can be very useful, because it reduces the waiting time as experienced by the end-user.
Nice examples of progressive data transfer are raster images, which are first
presented relatively fast in a coarse manner and then refined when the user waits
Marian de Vries and Peter van Oosterom
and edges (the GAP-edge forest). Faces and edges are assigned an importance range
(LOD range) for which they are valid. This set of tree-like structures is used when
the data is requested by a Web client to derive the geometry on-the-fly at the right
LOD, based on the importance values that were assigned during the generalization
process.
In the first part of this chapter we give a short overview of the tGAP structure. In
Sect. 5.2 the basic principles are described, followed by an explanation in Sect. 5.3
of how the generalization process builds a tGAP data set in a succession of steps.
In the second part of the chapter we explore how the tGAP data structures can be
used in a Web service/client environment. The focus is on two aspects: how the
tGAP structure can support progressive transfer of vector data from Web service to
client, and how adaptive zooming can be realized, preferably in small steps (‘smooth’
zooming). The relevant standards and protocols for vector data Web services are
discussed in Sect. 5.4: web feature service (WFS) and geography markup language
(GML). Section 5.5 explains how the tGAP structure can be used for progressive data
transfer and smooth zooming and proposes some necessary extensions to the current
standards (WFS and GML) in order to support vario-scale geo-information. Finally,
Sect. 5.6 concludes this chapter with a summary of the most important findings and
suggestions for further research.
5.1.1 Related Work
Vario-scale
Data structures supporting variable scale data sets are still very rare. There are a
number of data structures available for multiscale databases based on multiple representations (MRDBs) fixed number of scale (or resolution) intervals [7, 9, 25]. These
multiple representation data structures try to explicitly relate the objects at the different scale levels, in order to offer consistency during the use of the data. Most data
structures are intended to be used during the pan and zoom (in and out) operations
of a user, and in that sense multiscale data structures are already a serious improvement for interactive use as they do speed-up interaction and give a reasonable representation for a given LOD (scale). Drawbacks of the multiple representation data
structures are that they store redundant data (same coordinates, originating from the
same source) and that they support only a limited number of scale intervals.
Progressive Transfer
Another drawback of the multiple representation data structures is that they are not
suitable for progressive data transfer, as each scale interval requires its own (independent) graphic representation to be transferred. In a Web service/client context
progressive transfer can be very useful, because it reduces the waiting time as experienced by the end-user.
Nice examples of progressive data transfer are raster images, which are first
presented relatively fast in a coarse manner and then refined when the user waits