F. Bouille : The Use of G.I.S. in Petroleum Industry
247
graphical representation with images and texts appearing inside this structure, by
clicking on any component, using the new vizualization devices. Most problems
can be solved this way, combined with other tools (expert system, neural models,
etc .... ).
Largely developed in all fields of industry, particularly studied in Japan, fuzziness
is present everywhere in geographical phenomena; by approximation, in
coordinates for instance; by uncertainty, for instance concerning the possible
superimposition or intersection of networks; by subjectivity because of the opinion
that any geographer may express about the types he has captured and/or
processed; by multiplicity, because of the existence of several data instead of one
(where we expect only one), among which we cannot choose. Considering a
nuclear factory, we know that the effective places of components do not
correspond exactly to those mentioned on the plans; all information, even in such
a matter, are fuzzy, combining the four kinds of fuzziness: approximation,
uncertainty, subjectivity, multiplicity. Geography, like all activities, is deeply
involved in mastering fuzziness, even if some geographers have not realized that
the time has come to deal with this.
Though closely linked to the preceding topics, the quality is rather different, and is
limited or surrounded by fuzziness; geographers must evaluate the quality of the
data captured and stored; cartographic activity for instance may be compared to a
complete path inside a car factory; the concept of a flexible workshop may be
applied, as well as methods such as the KANBAN, nevertheless, with a difference:
the geographer uses the existing world and does not build it in a factory; but the
products he builds, such as maps, can be controlled by the same efficient methods.
Data quality evaluation is only just starting in GIS but will very soon find the same
place as in all other industries.
One of the most important features of the 0.-0. approach is self-extensibility; new
types may be defined owing to the basic types, by elementary schemes; it is then
very easy to build prototypes which can be used as a toy ("Iego" for instance) for
making copies, for assembling them, combining each prototype partially or
entirely; using prototypes such as "graph", "network", "isoline", "polyhedron", etc ...
immediately provides schemes composing any geographical structure, whatever
the theme (electricity, sewer, metro liner, geology, remote-sensing, etc ... ). It is the
first step to re-usability. Designing a geographical structure no longer requires
several months but only two or three days. Compared to the work needed for
building a relational model, such a method appears nearly instantaneous.
2.2 GIS: the features of the system
0.-0. databases are unlimited, shareable, protected, they can be built for storing
persistent abstract data types, and they can store simultaneously all kinds of
information: digital, textual, graphical, image, sound, etc .... The distinction
between metrical databases and topological databases appears particularly out-ofdate: using an 0.-0. structure shows that a class carries both its topology and its
metrics, and the term "database" looks obsolete. There are not several databases in
an 0.-0. persistent context; there is only the automated storage of all persistent
types, about which we need not know what device, what file, what table, etc ... are
used for ensuring this storage. An 0.-0. database supports all aspects of fuzziness
with no problem.
247
graphical representation with images and texts appearing inside this structure, by
clicking on any component, using the new vizualization devices. Most problems
can be solved this way, combined with other tools (expert system, neural models,
etc .... ).
Largely developed in all fields of industry, particularly studied in Japan, fuzziness
is present everywhere in geographical phenomena; by approximation, in
coordinates for instance; by uncertainty, for instance concerning the possible
superimposition or intersection of networks; by subjectivity because of the opinion
that any geographer may express about the types he has captured and/or
processed; by multiplicity, because of the existence of several data instead of one
(where we expect only one), among which we cannot choose. Considering a
nuclear factory, we know that the effective places of components do not
correspond exactly to those mentioned on the plans; all information, even in such
a matter, are fuzzy, combining the four kinds of fuzziness: approximation,
uncertainty, subjectivity, multiplicity. Geography, like all activities, is deeply
involved in mastering fuzziness, even if some geographers have not realized that
the time has come to deal with this.
Though closely linked to the preceding topics, the quality is rather different, and is
limited or surrounded by fuzziness; geographers must evaluate the quality of the
data captured and stored; cartographic activity for instance may be compared to a
complete path inside a car factory; the concept of a flexible workshop may be
applied, as well as methods such as the KANBAN, nevertheless, with a difference:
the geographer uses the existing world and does not build it in a factory; but the
products he builds, such as maps, can be controlled by the same efficient methods.
Data quality evaluation is only just starting in GIS but will very soon find the same
place as in all other industries.
One of the most important features of the 0.-0. approach is self-extensibility; new
types may be defined owing to the basic types, by elementary schemes; it is then
very easy to build prototypes which can be used as a toy ("Iego" for instance) for
making copies, for assembling them, combining each prototype partially or
entirely; using prototypes such as "graph", "network", "isoline", "polyhedron", etc ...
immediately provides schemes composing any geographical structure, whatever
the theme (electricity, sewer, metro liner, geology, remote-sensing, etc ... ). It is the
first step to re-usability. Designing a geographical structure no longer requires
several months but only two or three days. Compared to the work needed for
building a relational model, such a method appears nearly instantaneous.
2.2 GIS: the features of the system
0.-0. databases are unlimited, shareable, protected, they can be built for storing
persistent abstract data types, and they can store simultaneously all kinds of
information: digital, textual, graphical, image, sound, etc .... The distinction
between metrical databases and topological databases appears particularly out-ofdate: using an 0.-0. structure shows that a class carries both its topology and its
metrics, and the term "database" looks obsolete. There are not several databases in
an 0.-0. persistent context; there is only the automated storage of all persistent
types, about which we need not know what device, what file, what table, etc ... are
used for ensuring this storage. An 0.-0. database supports all aspects of fuzziness
with no problem.