In order to reference data sets on the Internet, DODS must use a naming convention which includes more information than the conventional pathname. Uniform Resource Locators (URL), as defined by the World Wide Web (WWW) project also serve that purpose. Because URLs are a de facto standard, adopting their syntax simplifies construction of DODS software, and eases integration of DODS with the emerging Internet information base.
DODS restricts the set of URLs it can use to a proper subset of the URLs acceptable to the WWW. This is done principally by stipulating that all data access takes place via one or more Common Gateway Interfaces (CGI) which extract data from the data set and return an experimental-type binary MIME document.
URLs, both as defined by the WWW project and in DODS's restricted sense, have a number of short comings. The URL notation does not carry information about the quality or longevity of the referenced resource. Furthermore, URLs are dependent on current naming conventions used on the Internet and thus lack `location independence'. However, the focus of DODS is in providing network accessibility to scientific data and not solving the problems associated with URLs as currently used by much of the existing Internet. By using URLs, DODS stands to benefit from the ongoing efforts of groups that are addressing the known deficiencies of URLs.