The NOAAServer Project: A WWW-Based Information Discovery and Retrieval Approach
Ernest Daddio
Anne O'Donnell
Wayne Brazille
National Oceanic and Atmospheric Administration
Environmental Information Services
1315 East-West Highway
Station 15548
Silver Spring, Maryland
Phone: 301-713-1262
Fax: 301-713-1249
edaddio@esdim.noaa.gov
edaddio@hpcc.noaa.gov
The NOAA Data and Information Server (NOAAServer) project was initiated in
FY95 under the sponsorship of the ESDIM Program to address the technical and
organizational issues associated with WWW-based on-line data and information
access. This project is a demonstration effort with several major goals:
(1) develop and implement an intuitive and unified information discovery and
retrieval mechanism across nationally distributed NOAA environmental
information systems, (2) provide access to a broad range of data and
information products spanning the spectrum from text, to point measurements,
to satellite imagery, (3) implement a common "look and feel" across
information servers to establish a corporate NOAA presence on the Internet,
(4) develop a virtual "one-stop shopping" capability for environmental
information discovery and retrieval, (5) promote the use of standards in
the dissemination of metadata and data products.
The NOAAServer concept is to guide the Internet-connected user to discovery
and retrieval of NOAA environmental data and information through a process
that takes him from a broad and potentially non-specific information
request, to a series of information discovery and refinement steps, to
previewing or browsing of data and information for applicability to the
specific user needs, and finally to retrieval or ordering of the information
product. The operational scenario is as follows. The user enters the
NOAAServer set of information systems initially through a keyword and
time/space domain query directed at a distributed database of metadata. The
result of the query is a group of data set titles displayed for him with
hypertext links to the remote servers containing the desired data or
information. The user may choose the appropriate data set to explore by
clicking on the data set title. He is then presented with a screen
generated by the remote server servicing that data set. This screen
provides a summary description of the data and presents the user with three
additional choices to refine his search or to acquire the data. The three
selections "Preview Data," "More Information," and "Obtain Data," are
hyper-linked within the server to additional screens that allow the user to
browse graphical data products, obtain more detailed text information on the
data, or be linked to the database for subsetting or data delivery,
respectively.
The networked system that has been implemented to date relies on the WWW
http protocol in a client-server model. Text searches are built on a Wide
Area Information Server (WAIS) search engine utilizing the Mosaic forms
capability to allow user entry of search parameters. A key element of the
implementation is that the metadata database distributed across servers.
This requires each of the participants to establish and maintain an indexed
WAIS database on their local server. An important feature of the metadata
search and retrieval is that information may be displayed by the user in any
of several standard formats including those prescribed by Federal mandate,
i.e., Federal Geographic Data Committee (FGDC) and Government Information
Locator Service (GILS). The distributed metadata databases may, in fact, be
in either NEDRES, Directory Interchange Format (DIF), or FGDC format; the
stored format is transparent to the user. To link the data set names to the
providers' servers, Universal Resource Locators (URL) are generated "on the
fly" to allow the user to point and click and to open a connection to the
remote data provider's server. Once connected to the provider's server the
user may browse a set of data products delivered to him in WWW standard
image format, such as GIF. These browse products are data set dependent
and may be generated dynamically according to user-prescribed criteria or
pre-generated by the provider and stored for retrieval. Alternatively,
the user may click on the "Obtain Data" selection and be connected to a
detailed inventory of the server's data. Within this screen, the user may
subset the data by time and geographic domain (within servers that provide
subsetting) through a WWW form and then retrieve the data on-line or have
it staged to a directory for FTP downloading. The user requiring more
details on the data prior to downloading or ordering, may click on the "More
Information" selection to view detailed text information on the data set.