1 Introduction 
Geographic Information Systems (GIS), has be- 
come and will grow to be a major application 
area of database management systems that rides 
on the information super-highway in the future. 
Essentially, it is a tool for simulating, modelling, 
analyzing and visualizing geographic features and 
geophysical phenomenon of the world around us. 
Its scope of application is unbounded. Typical 
application areas include: development of elec- 
tronic maps, geographic encyclopedia, natural re- 
source exploration and management, emergency 
dispatching, transportation, utility management, 
real estate management, city and community de- 
velopment, environmental protection, etc., [2]. 
GIS's applications impact all spheres of our daily 
lives. 
At the heart of a GIS system is the informa- 
tion base, i.e., a collection of databases that store 
the geometric definition of spatial features and 
objects and their related information relevant to 
the visualization process. These databases usu- 
ally are of different classes data. These may be 
classified by their data types into five principal 
types: 
Vector Data: The representation of spatial 
features by the definitions of their geomet- 
ric forms as points, lines, polygons, surfaces 
and solids. 
Raster Data: A representation of spatial fea- 
tures as digitized images. This is the typical 
format for still images, either arial photos or 
remote sensed images, and video - a conti- 
neous recorded digitized image of some geo- 
physical phenomenon. 
Structured Text: Organized text or at- 
tributes of spatial data that describe char- 
acteristics of spatial features. These are de- 
fined as records with fixed or variable length 
fields of text, symbols and icons. 
Free Text: Free formatted text in the form of 
documents. e.g., treaties, reports, land titles 
or deeds, etc. 
Knowledge Base: Recorded facts and 
derivation rules for logically deducing further 
facts and making logical inferences. 
Albeit not every GIS application requires all 
five classes of data-types, a GIS application nor- 
mally requires data from two or more of the above 
classifications. One problem normally encoun- 
tered in GIS today is that the available data is 
generally not in the format required by the user. 
Some processing and restructuring are required 
before it can be incorporated into the user's ap- 
plication environment. About seventy percent 
of the time to develop a GIS application, is de- 
voted to structuring, organizing and setting up 
the database. The rest of the time is divided be- 
tween developing the user interface (about 20%) 
and the analysis and display (about 10%). The 
latter activities are aided by the software tools 
and libraries provided by the GIS software. 
Naturally, the time and effort spent on refor- 
matting the data can be reduced if the data is 
already available in a format that is readily uti- 
lized. The phrase making data GIS ready has 
been coined for this effort. This has driven or- 
ganizations to take seriously, the development of 
standard exchange formats. Some of the well 
known initiatives include the development of DI- 
GEST [9], SAIF [20], and SDTS [22]. One still 
questions if the efforts being put into the devel- 
opment of such exchange formats are really justi- 
fied? 
Consider an alternative view of the prob- 
lem of making data GIS ready. We recognize 
that nearly all GIS software, e.g., ARC/INFO, 
SPANS, CARIS, MAPINFO, etc., retain spatial 
data in some proprietary data structure and then 
manage the related attribute information with a 
database management system (often, a relational 
DBMS such as INFO, Oracle, Ingres, etc.). The 
main reason for maintaining the spatial data in 
a proprietary format, it is argued, is for perfor- 
mance efficiency. Given that the spatial data 
can be maintained as a set of relational tables, 
it is clear then that the total information within 
the GIS can still be maintained by a database 
management system. If for performance reasons, 
a vendor requires, proprietary format to store 
the data, this can easily be retrieved from the 
database and restructured into the vendor spe- 
cific GIS software environment. Such an internal 
representation is normally irrelevant to the GIS 
users. 
Our main thesis in this paper then is that the 
problem of data interchange between the GISs is 
502 
really that c 
between dat: 
ternative vie 
is the motiv: 
of a Federat: 
System code 
The curre 
velopment is 
able from ag 
erational rea 
zational stru 
Development 
different org: 
to different 1 
out of date. 
data produci 
tion. For exe 
selected feat: 
must purcha 
that map she 
his applicatio 
avoid the cos 
digital inforn 
chased once 
when the dat 
in which diff 
vise their dat 
are very dyn: 
require them 
from the aut 
slow paced ii 
proliferation 
digital inforn 
curacy. 
Granting t] 
on the underl 
some of these 
providing an 
a distributed 
the scope of 1 
dressed by a 
system [3, 4, 
tem [3, 13. 1 
greater propc 
lems not easi 
Issues of Le 
nized thz 
predates 
nology. 
sets are 
DBMS.