104
Marian de Vries and Peter van Oosterom
For the implementation of line simplification based on BLG-trees a new data
type in GML would be useful. Client software then ‘knows’ how to react on receiving
BLG-tree geometry and standards-based line simplification routines in the client will
be enhanced.
For the same reason it would be useful to have standard attributes for the importance values of each feature (imp low and imp high), instead of having user-defined
ones.
5.5.3 Implications for Server and Client Software
In order to accomplish the progressive refinement when the geo-data is retrieved
from the WFS-R service an ‘order by’ expression (or functional equivalent) has to be
included in the WFS query to the data source. Necessary requirement is therefore that
the WFS service can retrieve the geo-objects from the data source in a sorted way.
A ‘Progressive Refinement’ WFS-R service, that serves data in order of importance
and in a certain importance range, does not need large extensions to WFS software.
What is extra in the WFS layer is an follows. (1) adding ‘order by’ to the queries sent
to the data source, and (2) adding importance selection to the filter conditions (either
spatial or non-spatial) that the user already has specified in the request to the WFS.
A WFS-R client must be able to hide the coarser geo-objects when the more detailed geo-objects are received and visualized and vice versa (during zooming out).
For polygons (with a solid color, not transparent) the painters algorithm takes care
of hiding the previously received objects. For (partly) transparent polygons and for
lines and points the visibility of the objects has to be manipulated by the client software, not only during zooming and panning, but also during progressive rendering
in the initial request. In case of line simplification also the topology-to-polygon construction has to be handled in the client.
5.6 Conclusions
In this chapter we presented a generalization approach based on offline generalization
that results in a variable scale, topological data structure (or rather a set of data structures): the tGAP structure. This approach has a number of advantages over multiscale/multirepresentation databases: there is no geometry redundancy, data consistency is therefore easier to maintain under updates; the structure is very fine-grained,
which makes smooth zooming possible; and the hierarchical (tree) structures enable
progressive transfer.
This chapter did illustrate the functioning of a vario-scale structure in a Web
service/client context. The tGAP structure is very well suited to a Web environment:
the client requirements are relatively low (almost no geometric processing of the data
at the client side) and progressive transfer of vector data is supported (allowing quick
feedback to the user). That the tGAP structure can be used for progressive transfer
Marian de Vries and Peter van Oosterom
For the implementation of line simplification based on BLG-trees a new data
type in GML would be useful. Client software then ‘knows’ how to react on receiving
BLG-tree geometry and standards-based line simplification routines in the client will
be enhanced.
For the same reason it would be useful to have standard attributes for the importance values of each feature (imp low and imp high), instead of having user-defined
ones.
5.5.3 Implications for Server and Client Software
In order to accomplish the progressive refinement when the geo-data is retrieved
from the WFS-R service an ‘order by’ expression (or functional equivalent) has to be
included in the WFS query to the data source. Necessary requirement is therefore that
the WFS service can retrieve the geo-objects from the data source in a sorted way.
A ‘Progressive Refinement’ WFS-R service, that serves data in order of importance
and in a certain importance range, does not need large extensions to WFS software.
What is extra in the WFS layer is an follows. (1) adding ‘order by’ to the queries sent
to the data source, and (2) adding importance selection to the filter conditions (either
spatial or non-spatial) that the user already has specified in the request to the WFS.
A WFS-R client must be able to hide the coarser geo-objects when the more detailed geo-objects are received and visualized and vice versa (during zooming out).
For polygons (with a solid color, not transparent) the painters algorithm takes care
of hiding the previously received objects. For (partly) transparent polygons and for
lines and points the visibility of the objects has to be manipulated by the client software, not only during zooming and panning, but also during progressive rendering
in the initial request. In case of line simplification also the topology-to-polygon construction has to be handled in the client.
5.6 Conclusions
In this chapter we presented a generalization approach based on offline generalization
that results in a variable scale, topological data structure (or rather a set of data structures): the tGAP structure. This approach has a number of advantages over multiscale/multirepresentation databases: there is no geometry redundancy, data consistency is therefore easier to maintain under updates; the structure is very fine-grained,
which makes smooth zooming possible; and the hierarchical (tree) structures enable
progressive transfer.
This chapter did illustrate the functioning of a vario-scale structure in a Web
service/client context. The tGAP structure is very well suited to a Web environment:
the client requirements are relatively low (almost no geometric processing of the data
at the client side) and progressive transfer of vector data is supported (allowing quick
feedback to the user). That the tGAP structure can be used for progressive transfer