;IES
example, may fall
a. Interestingly,
ng and frequently
this paper we will
yes and associated
“orporation which
nts and its impact
lems are not the
ler methods for
e conducive to
than the more
key is to look at
data are linked
nkage employed
| determine the
ires will require
lly don’t worry
cally coincident
pt of a feature is
"s interpretation,
ure. A user of
lassification, for
up of similarly
n. By focussing
n the same grid,
op is coincident
e two bands of
ed for each and
ss of the user's
vithin each band
ve could say that
r's interpretation
form.
special handling
features simply
e data structure
ept of a feature.
an be added with
soon as the user
attempts to extract the perceived features as vector
polygons, however, he or she enters a different
world of data linkage.
1.2 Implied Attributes
The simplest form of attribution for vector
objects is based on feature groupings usually
called levels or layers. A level is reserved for
each real-world feature to be discriminated in the
data. Early efforts at automated mapping with
Intergraph systems, for example, frequently used
this strategy. When the levels are known only by
a unique number, the user usually maintains a
reference document which explains the various
level assignments. AutoCAD-based systems have
some advantage insofar as the level names are
user-defined and therefore self-documenting.
When two features are geographically
coincident in this system, they must be
represented by redundant graphic objects placed in
each level corresponding to the coincident real-
world features (See Figure 2).
The advantage of this system is its
simplicity. Since there is no external database,
this is an ideal system of attribution when a single
attribute (feature level) can be used to discriminate
all features. Translation to other spatial
information systems is relatively painless so long
as the feature level survives translation.
Intergraph and AutoCAD formats have emerged
as de facto standards for this system of attribution
and most GIS software will import and export
data in these formats preserving the feature level
Real-World B
Features A C
PR nt ri ve us
Graphic ‘Level 1 i
Objects Yevel2 0
EN BEC Lei boo luu. 2.
User
Lookup Table
Level 1 = A
Level 2 = B
Level 3 = C
Figure 2. Implied Attributes
as a user attribute.
The disadvantage of this system is the
potential for asynchronization between coincident
features on different levels. Coincident features
are usually copied from one level to another at the
time of initial automation, but during the lifetime
of the data, as it receives graphic updates, there is
always the potential that one feature will be
changed, while its coincident partner is ignored.
Procedures must be implemented which link
coincident features in such a way that if the
system cannot automatically synchronize the
geometry between features, then it must at least
alert the user to the existence of coincident
partners when a graphic change is attempted.
These procedures can require low-level application
programming and/or additional maintenance labor
costs which erode the savings anticipated by
choosing the relatively simple system of implied
attribution.
1.3 Selective User Linkage
Many mapping systems store features as
vector objects with selective linkage to an attribute
database. Most un-enhanced CAD-based
applications today fall into this category. A
selective linkage between graphic objects and their
attribute records allows the user to attach any
number of attribute records to a single graphic
object. Real-world features which are
geographically coincident, therefore, can find
representation as distinct attribute records sharing
a common graphic object (See Figure 3).
Selective linkage is an efficient method for
providing intelligence to a limited number of
intelligent features, especially when they may be
surrounded by many non-intelligent, or "dumb"
features. Examples are common in the AM/FM
world where mapping systems are used to track
specific "outside plant" assets surrounded by
dumb representations of geographic landmarks.
One drawback to the selective-linkage
approach, however, is that it does require effort,
usually human labor, to establish and maintain the
individual links between graphic objects and
attribute records. This effort can be cost-
prohibitive if the data are subject to frequent
updates, especially graphic updates which are
often multiplied when the spatial relationship
between features is significant. For this reason,
the majority of GIS applications today employ a
system of universal automatic linkage.