CoastGls'99: geomatics and coastal environment
Wider contexts
These discussions and issues also mirror events taking place in the evolution of GIS generally. The whole field of GIS appears to be undergoing rapid change and evolution at the moment, both technologically
and institutionally, though at least sorne of these changes almost
suggest a return to earlier ways of doing things (in sorne cases, more
efficiently in the light of technical imptovements) rather than the introduction of radically new methods and techniques. For example, in the
early days of A rc/Info and other pioneering proprietary systems, GIS software was often packaged as a "toolbox", with a wide array of utilities
and functions provided with the system, but as separate modules, so
that the user could select the specific tools required for a particular
task and apply these as required. Later on came the more integrated
approach, where aIl possible tools and functions were provided within
a single package, with a common interface, whether they were required
or not. This meant that users were inevitably stuck with the weaknesses of whichever GIS product they had selected, as weIl as its strengths.
For coastal GIS this has been particularly relevant, since (as is often
asserted) almost aIl commercially-developed GIS have been designed
and optimised for handling terres trial rather than coastal or marine
data; thus, the coastal GIS user was more-or-Iess forced to use tools
and techniques that were inherently weak for the specific application(s)
concerned.
In recent years, we have seen a move back towards the toolbox approach
to GIS, this time based on a "plug and play" concept. The bricks in a
child's "Lego" set offer a good analogy: in a "Lego" set, we find many
general-purpose bricks, and also a smaller number of more specialized
components. Whether generic or more specific, each piece has been
designed to facilitate easy interconnection and, in this way, can be
assembled so as to create whatever model the child wishes. Similarly,
emerging standards of interoperability are encouraging software developers to create application-specific modules that "plug into" a core set
of generic GIS functions. Thanks to this new approach, people will
likely soon be able to select and assemble their own custom-built geoprocessing system from tools that best suit their specific needs and applications (in practice this assembly will often be undertaken for the
end-user by specialist consultants or other intermediaries who are expert
in system selection and interfacing). Sorne of the elements in these custombuilt systems may be unorthodox by today's standards but, by integrating
them within a single coherent framework, we will be able to harness
the synergy of their interaction, and mrn these separate tools (including
many elements of existing GIS) into dedicated and specialized coastal
information systems. A number of the presentations made at this
meeting have reflected this evolution, and we already can see clear
benefits, in terms of utility and flexibility of such approaches.
308
Wider contexts
These discussions and issues also mirror events taking place in the evolution of GIS generally. The whole field of GIS appears to be undergoing rapid change and evolution at the moment, both technologically
and institutionally, though at least sorne of these changes almost
suggest a return to earlier ways of doing things (in sorne cases, more
efficiently in the light of technical imptovements) rather than the introduction of radically new methods and techniques. For example, in the
early days of A rc/Info and other pioneering proprietary systems, GIS software was often packaged as a "toolbox", with a wide array of utilities
and functions provided with the system, but as separate modules, so
that the user could select the specific tools required for a particular
task and apply these as required. Later on came the more integrated
approach, where aIl possible tools and functions were provided within
a single package, with a common interface, whether they were required
or not. This meant that users were inevitably stuck with the weaknesses of whichever GIS product they had selected, as weIl as its strengths.
For coastal GIS this has been particularly relevant, since (as is often
asserted) almost aIl commercially-developed GIS have been designed
and optimised for handling terres trial rather than coastal or marine
data; thus, the coastal GIS user was more-or-Iess forced to use tools
and techniques that were inherently weak for the specific application(s)
concerned.
In recent years, we have seen a move back towards the toolbox approach
to GIS, this time based on a "plug and play" concept. The bricks in a
child's "Lego" set offer a good analogy: in a "Lego" set, we find many
general-purpose bricks, and also a smaller number of more specialized
components. Whether generic or more specific, each piece has been
designed to facilitate easy interconnection and, in this way, can be
assembled so as to create whatever model the child wishes. Similarly,
emerging standards of interoperability are encouraging software developers to create application-specific modules that "plug into" a core set
of generic GIS functions. Thanks to this new approach, people will
likely soon be able to select and assemble their own custom-built geoprocessing system from tools that best suit their specific needs and applications (in practice this assembly will often be undertaken for the
end-user by specialist consultants or other intermediaries who are expert
in system selection and interfacing). Sorne of the elements in these custombuilt systems may be unorthodox by today's standards but, by integrating
them within a single coherent framework, we will be able to harness
the synergy of their interaction, and mrn these separate tools (including
many elements of existing GIS) into dedicated and specialized coastal
information systems. A number of the presentations made at this
meeting have reflected this evolution, and we already can see clear
benefits, in terms of utility and flexibility of such approaches.
308