129 
The map layer “region of DTM source code” may be 
described in GRASS by the following files: 
• a file in dig_ascii directory, containing an header 
section which identifies the type of information ( title 
of the map, scale of the original source map,...) and a 
section with the data. In the second part of the file 
each area or line feature of the map is described by 
means of a code (A stands for area and L for line), 
the number of points corresponding to the geometric 
description of the feature, the 9, X coordinates of the 
vertices. For example in our case the header section 
and a piece of data section is shown in Figure 4. 
• A file in the dig_att directory, in which the attributes 
associated to the area or line feature are described 
(one for each map feature). 
In each record of this file the feature type, the X, 9, 
coordinates of a centroid and a category number 
(attribute) corresponding to the feature are stored. 
The example corresponding to the case we are 
examining is given in Figure 5. 
ORGANIZATION: IGES 
DIGIT DATE: 22/02/1999 
DIGIT NAME: 
MAP NAME: 
MAP DATE: 
OTHER INFO: 
MAP SCALE: 
ZONE: 
WEST EDGE: 
EAST EDGE: 
SOUTH EDGE: 
NORTH EDGE: 
MAP TRESH: 
VERTI 
M. A. BROVELLI 
CODEDTM1 
1994 
DTM VECTORIZED 
25000 
32 
4 
21 
34 
49 
0.0 
A 178 
46.37 7.71 
46.63 7.71 
46.63 7.44 
46.56 7.44 
Fig. 4 - The file describing the area features 
corresponding to the map in Fig. 2. 
Fig. 5 - The file describing the attribute “source code” 
related to the sources of the Italian DTM in Fig. 2. 
Now, if we would like to transpose these information in 
terms of RDBMS tables, the basic relations which allow 
for the geometrical description of our information region 
layer are described in Tables 7,8,9 and 10. 
Pointjd 
<P 
X 
1 
46.37 
9.12 
2 
46.37 
10.71 
3 
47.63 
10.71 
4 
47.63 
9.44 
5 
47.40 
9.44 
6 
47.40 
7.60 
178 
46.20 
9.56 
179 
46.37 
9.60 
Table 7 - “Point” relation 
Arcjd 
Pointjd 
Point_sequence _number 
L1 
1 
26 
L2 
2 
1 
L2 
3 
2 
L3 
3 
1 
L3 
4 
2 
L3 
5 
3 
L3 
6 
4 
L4 
6 
1 
L4 
L5 
L5 
179 
9 
Table 8 - “Arc” relation 
Polygonjd 
Arcjd 
A6 
L2 
A6 
L3 
A6 
L4 
A6 
L5 
Table 9 - “Polygon” relation 
Regionjd 
Polygonjd 
DTMJd 
R6 
A6 
6 
... 
Table 10 - “DTM source region” relation 
This approach howeveris not followed because of the loss 
of storage space, due to 
• repetition of the Arcjd in the Arc relation; 
• repetition of the Polygonjd in the polygon relation; 
• need to store also the Point_sequence_number in the 
Arc relation. 
The resulting “long thin” tables impose the creation of 
large indexes, which are not efficient in rapidly retrieving 
the data.