246
Accessing Geoenvironmental Data: From On-line to Expert Systems
1
INTRODUCTION
The petroleum industry is largely diversified, and managing the produced data
first requires knowledge of many rather different topics, but always centred on two
purposes: exploration and production. It needs various studies such as geophysics
and diagraphies, together with many other topics dealing with the 2-D or 3-D (and
even 4-D) coordinates; in fact, they merely deal with the geography, with metrical
properties as well as topological relationships. That is why the GIS (Geographic
Information Systems) are so important in oil exploration and production activities.
2
THE MAIN TRENDS IN THE GIS
The concept of GIS, appeared in the 1970s but was really accepted fifteen years
later. Many tools have been developed during the last decade, generally based
upon concepts designed a long time ago, all stemming from "brainware"
(Kitagawa, 1974), with implications on software and hardware. New ways have
been introduced for considering geographical problems, new ways of
programming them have been developed, new integrated systems have been
realized, and new devices have been built, revolutionizing the way of using a
computer; and finally, new built-in geographical components have been prepared,
introducing the concept of re-usability of geographical types (Bouille, 1994d).
2.1 New concepts in GIS
These concepts are essential in GIS, where they appeared in 1976. If everybody
agrees with multiple inheritance, the use of methods, possible communication
through messages, etc ... , some other ideas introduced eighteen years ago have not
been completely understood by all: -persistency, -graphical representation of
types, -fuzziness, ...
The most important aspect concerns persistency: any geographical type must be
created, updated, used as if it was resident in a core memory, and it must be
automatically stored in a very large hidden database, without any human action;
the user will find this type the next time he is connected, as if it had never left the
core memory. This point is closely linked to 0.-0. databases. Working with
geographical types requires persistency, because it provides the most simplicity for
the user and the greatest efficiency for the system.
Some authors trying to describe a phenomenon with a programming language
speak about "data representation". It is a very poor representation; for a complex
phenomenon requiring 3000 abstract data types, 1500 pages are sometimes
necessary.
Associating graphical representation with types, independent of any programming
language, is a powerful tool for communication between human beings, even if
they come from several different countries, because it is a common language; it is
a way too of avoiding classical programming in most basic operations, and it
explains the solution to a problem, no matter how complex, by a path inside the
graphical knowledge structure. The next step, more recent, is the use of 3-D
Accessing Geoenvironmental Data: From On-line to Expert Systems
1
INTRODUCTION
The petroleum industry is largely diversified, and managing the produced data
first requires knowledge of many rather different topics, but always centred on two
purposes: exploration and production. It needs various studies such as geophysics
and diagraphies, together with many other topics dealing with the 2-D or 3-D (and
even 4-D) coordinates; in fact, they merely deal with the geography, with metrical
properties as well as topological relationships. That is why the GIS (Geographic
Information Systems) are so important in oil exploration and production activities.
2
THE MAIN TRENDS IN THE GIS
The concept of GIS, appeared in the 1970s but was really accepted fifteen years
later. Many tools have been developed during the last decade, generally based
upon concepts designed a long time ago, all stemming from "brainware"
(Kitagawa, 1974), with implications on software and hardware. New ways have
been introduced for considering geographical problems, new ways of
programming them have been developed, new integrated systems have been
realized, and new devices have been built, revolutionizing the way of using a
computer; and finally, new built-in geographical components have been prepared,
introducing the concept of re-usability of geographical types (Bouille, 1994d).
2.1 New concepts in GIS
These concepts are essential in GIS, where they appeared in 1976. If everybody
agrees with multiple inheritance, the use of methods, possible communication
through messages, etc ... , some other ideas introduced eighteen years ago have not
been completely understood by all: -persistency, -graphical representation of
types, -fuzziness, ...
The most important aspect concerns persistency: any geographical type must be
created, updated, used as if it was resident in a core memory, and it must be
automatically stored in a very large hidden database, without any human action;
the user will find this type the next time he is connected, as if it had never left the
core memory. This point is closely linked to 0.-0. databases. Working with
geographical types requires persistency, because it provides the most simplicity for
the user and the greatest efficiency for the system.
Some authors trying to describe a phenomenon with a programming language
speak about "data representation". It is a very poor representation; for a complex
phenomenon requiring 3000 abstract data types, 1500 pages are sometimes
necessary.
Associating graphical representation with types, independent of any programming
language, is a powerful tool for communication between human beings, even if
they come from several different countries, because it is a common language; it is
a way too of avoiding classical programming in most basic operations, and it
explains the solution to a problem, no matter how complex, by a path inside the
graphical knowledge structure. The next step, more recent, is the use of 3-D