NOTE: The DODS and OPenDAP projects have used the OPeNDAP FreeForm ND Data Handler to present a catalog
of data files to the world as a single dataset. In many ways this was a very successful system, providing catalogs for multi-granule datasets that could be searched by date and time. However, the OPeNDAP project has decided (winter 2006) to adopt the THREDDS xml-based catalog system developed at Unidata, Inc. The remainder of this chapter describes the 'file servers' that can be built using the FreeForm data handler. Even though we feel it's best to adopt the THREDDS catalogs, there are good reasons to keep existing catalog servers running and to build new catalogs as a stop-gap measure to support existing client software.
Normally, in the OPeNDAP argot, a "dataset" is contained in a single file on a disk. However, this paradigm is often broken by large datasets that may contain many thousands or tens of thousands of data files. The OPeNDAP file server is a way to make these dicrete datasets appear to be a single large dataset.
The OPeNDAP file server is an OPeNDAP server that returns a URL or set of URLs in response to a query containing selection variables. For example, a dataset organized by date and geographic location might provide a file server that allowed you to query the dataset with a range of dates and longitudes. This fileserver would return a list of one or more URLs corresponding to files within that dataset that fell within the given range.