78 
Initial date 
Final date 
File format 
File location 
Comment 
Relations. 
These values are organized in a number of different 
relational tables depending upon the specific "entity type". 
For instance, for an entity of type "municipality" the 
"position" is in fact a collection of point coordinates 
representing its borders. For a measurement station the 
position slot is only a pair of coordinates. The same is true 
for the slot "relations" which defines the entities that are 
connected in some way to the one at hand (for instance, a 
lake has a relation with the inflowing river). 
Obviously, several different data sets may belong to the 
same entity and they may also belong to different data 
categories. For instance, a measurement station may 
collect flow data as well as air temperature or oxygen 
concentration. 
The “comment” field can be used to supply all additional 
information to the data set, such as methods of measure, 
data reliability, code for missing data, etc. 
The entity type is the only attribute that must be chosen 
within a specified value set: municipalities, mountain 
communities, parks, lakes, rivers, measurement stations. 
Extensions of this set imply a revision of the system, while 
for all the other sets (for instance, data categories or file 
formats) they can be easily expanded by the system 
administrator since their possible values are also stored in 
specific database tables. 
The definition and use of the attribute "relations" requires 
some more details. The entity types are organized in a 
hierarchy, which, together with the geographical location, 
automatically defines a certain number of relations 
between entities. Such a hierarchy is sketched in figure 4. 
For instance, a mountain community is automatically 
related to the municipalities that constitute it or a river is 
related to the park through which it flows. However, the 
entity at a higher level in the hierarchy are related with all 
the entity at a lower level, while the reverse it is not true. 
Entities at the same level are related only to those 
adjacent or included. The system administrator can also 
force other relations in specific cases. 
The existence of such relations among entities is of great 
importance for data retrieval. If a user is looking for air 
quality data in a certain community, he or she will be able 
to see all those collected by the individual municipalities, 
even if they are not directly in relation with the community 
as a whole. On the contrary, if the user looks specifically 
for the data in a certain municipality, that means he or she 
is not interested in the rest of the community and thus 
such data will not be provided. Note that air quality data 
are normally associated to measurement stations, but 
they can be found looking for a municipality since they 
have a relation of inclusion with it. 
4. SYSTEM ARCHITECTURE 
The architecture of the system is very simple and based 
on two different figures: users and administrator (fig. 5). 
Fig. 5 - Structure of the system 
Each of them has dedicated programs either in the client 
(the users) or in the server (both the users and the 
administrator) side of the computer implementation. 
The user can search for environmental data sets by 
preparing a query to the metadata base using the 
interface. He can search by selecting a geographical 
position (within a given distance from a point, whose 
coordinates are either entered numerically or by clicking 
the mouse on the map), by data categories, by time 
period or by a combination of these attributes. He can 
preview the data.sets found and decide whether or not to 
download them. He or she can also submit new data sets 
to the system. For this purpose, a frame with the 
metadata specification must be filled and the file 
containing the data set is uploaded to the server. Finally, 
the user can also suggest the creation of a new 
measurement station, by defining its position and name. 
Both these last possibilities require the intervention of the 
administrator. He is notified of the user’s request, must 
analyze the information submitted by the user and control 
that all the necessary definitions have been correctly 
supplied. Obviously, he is not responsible for the 
correctness or completeness of the data, which are 
entirely under control of the data provider. It is, on the 
contrary, a task of the administrator to check if the 
relations automatically established by the system for the 
new data set are correct or if other should be added.