80
Michela Bertolotto
• Geometry and Topology Integration: Approaches based on line simplification
are applicable only for linear data sets and do not take into account other
types of generalization operators (such as those involving changes in topology);
vice versa, approaches based on topological operators do not support geometric
changes. There is a need for the development of approaches that fully support
the integration of topological and geometric operators. Semantic aspects should
also be integrated.
• Data Consistency Control: Many existing algorithms do not implicitly preserve
consistency and still require a posteriori checks; automated mechanisms to guarantee this property must be studied (see also Chap. 8 for further discussions
about consistency between vector maps).
• Data Storage and Multiple Representations: Efficient hierarchical data structures for storing and manipulating different map versions are needed to facilitate
transmission. These should avoid duplication and re-transmission of data and include explicit links for effective navigation and querying across levels. Another
important challenge when using multiple representations consists of providing
a suitable answer to the question: “how many and what levels should be precomputed?” The choice of what representations need to be stored is often both
application- and data-dependent. This problem poses limitations and should be
further studied (see also Chap. 5 for further discussions on the use of multiple
representations for Web applications).
• Efficient Spatial Indexing Mechanisms: With available vector spatial data sets
becoming larger and larger, there is a need to efficiently index this data for fast
access and retrieval. Many of the system prototypes proposed (e.g. [8]) did not
consider this aspect. Others, that rely on commercially available spatial application servers (e.g. [51]) exploit the indexing functionality provided. However,
a full evaluation of how these indexes influence the performance has not been
conducted and constitutes an interesting research direction.
• Accuracy Control: A full study on evaluating the accuracy of the data exchanged
and how essential it is in the different application domains would also represent
an interesting study in the context of progressive data exchange.
From a more practical (implementation-oriented) point of view, further developments and decisions are required in relation to the following aspects:
• Definition of a Transmission Technique: This should be efficient (reduce users’
waiting time). Studies should be conducted on deciding, for example, whether to
segment transmitted data sets so as to transmit smaller files. However, this might
further complicate operations for topology reconstruction.
• Development of algorithms to efficiently compile transmitted increments into a
topologically consistent format for the reconstruction of intermediate levels of
detail. This should be facilitated by the use of vertical links.
• Development of graphic interfaces and dynamic visualization tools for displaying different map versions corresponding to different levels of detail (during
transmission).
Michela Bertolotto
• Geometry and Topology Integration: Approaches based on line simplification
are applicable only for linear data sets and do not take into account other
types of generalization operators (such as those involving changes in topology);
vice versa, approaches based on topological operators do not support geometric
changes. There is a need for the development of approaches that fully support
the integration of topological and geometric operators. Semantic aspects should
also be integrated.
• Data Consistency Control: Many existing algorithms do not implicitly preserve
consistency and still require a posteriori checks; automated mechanisms to guarantee this property must be studied (see also Chap. 8 for further discussions
about consistency between vector maps).
• Data Storage and Multiple Representations: Efficient hierarchical data structures for storing and manipulating different map versions are needed to facilitate
transmission. These should avoid duplication and re-transmission of data and include explicit links for effective navigation and querying across levels. Another
important challenge when using multiple representations consists of providing
a suitable answer to the question: “how many and what levels should be precomputed?” The choice of what representations need to be stored is often both
application- and data-dependent. This problem poses limitations and should be
further studied (see also Chap. 5 for further discussions on the use of multiple
representations for Web applications).
• Efficient Spatial Indexing Mechanisms: With available vector spatial data sets
becoming larger and larger, there is a need to efficiently index this data for fast
access and retrieval. Many of the system prototypes proposed (e.g. [8]) did not
consider this aspect. Others, that rely on commercially available spatial application servers (e.g. [51]) exploit the indexing functionality provided. However,
a full evaluation of how these indexes influence the performance has not been
conducted and constitutes an interesting research direction.
• Accuracy Control: A full study on evaluating the accuracy of the data exchanged
and how essential it is in the different application domains would also represent
an interesting study in the context of progressive data exchange.
From a more practical (implementation-oriented) point of view, further developments and decisions are required in relation to the following aspects:
• Definition of a Transmission Technique: This should be efficient (reduce users’
waiting time). Studies should be conducted on deciding, for example, whether to
segment transmitted data sets so as to transmit smaller files. However, this might
further complicate operations for topology reconstruction.
• Development of algorithms to efficiently compile transmitted increments into a
topologically consistent format for the reconstruction of intermediate levels of
detail. This should be facilitated by the use of vertical links.
• Development of graphic interfaces and dynamic visualization tools for displaying different map versions corresponding to different levels of detail (during
transmission).