226 Alian R. Robinson and Jurgen Sellschopp
A distributed data system for REA has to be better organized than ordinary information on the Internet. It is unacceptable for participants in an REA effort to regularly search through the offers of potential data providers and find out whether new
information has shown up. There must be an element in the REA organization,
called the Fusion Center, that either receives an REA data set as soon as it is generated, or is informed about its existence and the location from where it can be downloaded. The Fusion Center will set up an inventory of all existing data with pointers
to the respective files. It generates maps, designs web pages and provides appropriate data search algorithms. Occasionally the Fusion Center can adjust data formats
to a common standard. Data originators remain however responsible for data integrity. The Fusion Center provides service to participating groups who for product
generation rely on other people's data, it creates and maintains an archive of contributions, and it finally places results to be browsed by customers.
11.7.4 Simple aRd complex systems for special purposes
The complexity of the distributed data management system and fusion center
depends on the size of the REA operation. In the most simple case of a single platform for measurements and modeling, data colIection on a computer with a clear
directory structure and a few pages written in hypertext markup language (html)
for descriptions and cross-links between data types would do. In a complex REA
survey, data are physically organized relative to the different data providers. They
keep their contributions together, either on an own Internet server or in a directory
at the fusion center. The fusion center then creates tables that would combine links
to a certain data type regardless of the originator. The distributed data archive thus
gets a double structure: the physical placement of files and the menu-guided access
to data can differ significantly.
In an REA operation, it is often required to establish a back-up for the data server
or to save the complete server contents on a CD-ROM. This should be facilitated
by a simple information format and server structure. Some of the modern techniques, that are widely used on Internet servers such as java scripts and database
queries, are counterproductive in that respect.
Common formats for alI data delivered for REA, though desirable, are unlikely
accomplished. But care must be taken that data files are compatible between computer systems. Standard tools under the most frequently used operation systems, at
least MS Windows and Unix, must be able to digest alI files. Plain ASCII files with
a simple structure are preferable. If binary data are more appropriate, platform
independent formats such as Netcdf or Matlab save sets are recommended. For text
and image data, standard file types of the world wide web (Internet) are preferred
including the platform-independent document format (pdf) for vector graphics and
formatted documents. Postscript (ps) format is also acceptable for documents that
are anticipated for high quality printouts rather than for screen display.
Files generated with a typical office tool such as a spread sheet, presentation
graphic or document writer, would require compatible software releases for ingestion, that are unlikely available under Unix and may not be present under Win-
A distributed data system for REA has to be better organized than ordinary information on the Internet. It is unacceptable for participants in an REA effort to regularly search through the offers of potential data providers and find out whether new
information has shown up. There must be an element in the REA organization,
called the Fusion Center, that either receives an REA data set as soon as it is generated, or is informed about its existence and the location from where it can be downloaded. The Fusion Center will set up an inventory of all existing data with pointers
to the respective files. It generates maps, designs web pages and provides appropriate data search algorithms. Occasionally the Fusion Center can adjust data formats
to a common standard. Data originators remain however responsible for data integrity. The Fusion Center provides service to participating groups who for product
generation rely on other people's data, it creates and maintains an archive of contributions, and it finally places results to be browsed by customers.
11.7.4 Simple aRd complex systems for special purposes
The complexity of the distributed data management system and fusion center
depends on the size of the REA operation. In the most simple case of a single platform for measurements and modeling, data colIection on a computer with a clear
directory structure and a few pages written in hypertext markup language (html)
for descriptions and cross-links between data types would do. In a complex REA
survey, data are physically organized relative to the different data providers. They
keep their contributions together, either on an own Internet server or in a directory
at the fusion center. The fusion center then creates tables that would combine links
to a certain data type regardless of the originator. The distributed data archive thus
gets a double structure: the physical placement of files and the menu-guided access
to data can differ significantly.
In an REA operation, it is often required to establish a back-up for the data server
or to save the complete server contents on a CD-ROM. This should be facilitated
by a simple information format and server structure. Some of the modern techniques, that are widely used on Internet servers such as java scripts and database
queries, are counterproductive in that respect.
Common formats for alI data delivered for REA, though desirable, are unlikely
accomplished. But care must be taken that data files are compatible between computer systems. Standard tools under the most frequently used operation systems, at
least MS Windows and Unix, must be able to digest alI files. Plain ASCII files with
a simple structure are preferable. If binary data are more appropriate, platform
independent formats such as Netcdf or Matlab save sets are recommended. For text
and image data, standard file types of the world wide web (Internet) are preferred
including the platform-independent document format (pdf) for vector graphics and
formatted documents. Postscript (ps) format is also acceptable for documents that
are anticipated for high quality printouts rather than for screen display.
Files generated with a typical office tool such as a spread sheet, presentation
graphic or document writer, would require compatible software releases for ingestion, that are unlikely available under Unix and may not be present under Win-