The Distributed Oceanographic Data System:

Requirements for Data Analysis Applications


Access to distributed remote data alters both the scientist's use of data and expectations of analysis software. Access to remote data is fundamentally different from access to local data because the potential pool of resources is vastly larger. When confronted with this largesse, users requirements for data analysis software changes. Because users do not control remote data resources, they need ways to search among all available data sets for a particular type of information. These searches are different than the traditional database search because users, not knowing precisely what data exists, conduct their search in an iterative manner, asking repeatedly more specific questions. Users also want to select data from a data set in ways that suite their own work; this often means that remotely accessible data sets must support a more general query mechanism than those accessible only locally. Finally, users expect to be able to operate on several remote data sets at once, combining these different data in a single display. Within The Distributed Oceanographic Data System, we have had unique experience with the new requirements scientists place on data analysis applications because DODS provides a pathway for existing analysis programs to be extended so that they can access remote data resources.


James Gallagher