The segment node format of DCAD will have inherent advantages over the 
polygon format currently used for DFAD data storage. The polygon format 
does not identify nodes of feature segments. The whole feature (Fig. 3A) can 
be identified only as a segment. Thus, DFAD features sharing a common 
boundary were overlapped to insure coverage. This procedure was developed 
because, at that time, only manual digitizers were available for data collection 
and the capability to digitize a single common boundary for adjacent features 
was impractical. Manual overlap could be eliminated in off-line post processing 
of the digital data but often resulted in the subsequent computer generation of 
overlap slivers or fictitious gaps in the digital data and additional processing 
was required to eliminate them, where possible. 
The line segment noded format would eliminate this problem, avoid double 
storage or common boundaries, simplify update and correction, and be 
responsive to thinning and generalization algorithms. 
It is feasible to develop a storage format designed to accommodate these 
requirements. Such a design would include the following. 
1. Header information to contain descriptive information or attributes of the 
data set. 
2. Data sets to accommodate feature and line segment information. 
a. Feature information (TABLE 3) identifies the unique features which 
make up the data set. Specifics on feature data content would depend upon the 
product specification (Fig. 5). 
b. Line segment information (TABLE 3) contains the actual coordinate 
strings for the segments which make up each unique feature. All features 
which include a segment would be so identified. 
3. Free text information to further define or uniquely qualify data as required. 
POLYGON 
LINE SEGMENT 
FEATURE A: SEGMENT 1 
FEATURE B: SEGMENT 2 
FEATURE C: SEGMENT 3 
FEATURE A: SEGMENTS 1, 2 
FEATURE B: SEGMENTS 2, 3, 4, 5 
FEATURE B: SEGMENTS 4, 6 
A 
e 
Figure 3. Data Base Formats