In dynamic map handling, there are limited memory resources in the client devices, but remote access to the server can be assumed. Maps are obtained on-the-demand from the client device when available, or automatically through a GSM-network, for example. The idea is that only the necessary amount of data is transmitted. When the map fragments are transmitted in a compressed form, the current network has about enough bandwidth to handle the situation.

Dynamic map handling means also that the handling of the maps is invisible to the user: the maps just appear on the screen when needed without any actions by the user. Technically speaking the device indicates the present location to the map handling system, and requests to show a piece of map on the display. If the map does not exist in the memory, request is sent to the server via network. It is a matter of the application to define the logic to automatize the selection of scale and other parameters of the map. The server the looks for a proper map sheet from the database and sends the desired piece of the map to the client in a compressed form.

The map handling system must be simple but powerful enough to handle a large number of small map fragments from a very large area (Figure 3). The maps can be very fragmented and, therefore, the file and memory use must be taken care efficiently. If all memory area is filled with small map images, the memory handling can become a problem. In the REALMAP project, a suitable map format was created because it allows direct access to small map segments but still stores the segments from a single map sheet covering a large area in the same file. This solution allows flexible tailoring both to small hand-held devices with uniform shared memory area, and also to Laptop and PocketPC applications with specialized file handling system.

Figure 3: Dynamic map handling by composing map sheet from 
individual map segments.

Handling of the maps composed of small fragments requires sophisticated data structures supporting basic operations such as insert and delete of a map segment. The detail implementation of these operations depend on the application. In some application, the user may want to prioritize which maps he/she wants to preserve for certain, and which maps can be removed when not needed anymore. This can be automated to a certain extent on the basis of the location and history information. We can see to extremes in this user scheme: 
(1) fully dynamic,
(2) fully static.

In fully dynamic map handling the user has no control on the maps whatsoever. The maps just appear when needed. In fully static system, on the other hand, insertion and deletion of maps happens only by the explicit request by the user. Overall, the dynamic map handling requires flexibility from the file format. Insertion and deletions must be able to perform independent of the application and the logic that it will implement.

Dynamic map handling itself does not prevent intelligent application logic such as routing and guided navigation. The point is that if we must rely on the server to obtain map, we can utilize the server for performing the more complicated logic as well. In other words, the complex database systems can be located in the server whereas client stores the maps only as they are to be used. The server can also try to predict where the user is going to, and prepare the user with maps already beforehand. In this way, not only the map handling but also the utilization the network bandwidth would be dynamic in its nature.

<<<<goals

next>>>>