Figure 1: Typical Data Capture Systems
with the data (and not accessible) or are simply discarded. Thus, correlation of the science
data and its associated ancillary data is difficult.
Today’s data capture systems are generally designed to support multiple platforms and
instruments. These systems cannot function without data management, which in turn
depends upon rigorous standard methods of data description and identification. Scientific
analysis requires long-term access to the instrument data plus all of the supporting data.
Information can be defined as the communication or reception of knowledge or intelligence.
In the context of this document then, information transfer must include not only the science
data bits, but also the metadata needed for analysis of the data.
INFORMATION TRANSFER AND ARCHIVE CONCEPT
Central to the concept is the data product. It is a collection of data that can range from a
simple message to a large complex aggregation of data sets. A data product typically
contains not only the data (e.g. an image, a frame of instrument data) but all of the
supporting data (metadata) that is needed to identify and understand the data object. The
object here is to describe the abstract ways in which data products can be built from simpler
objects. The building process is essentially hierarchical. To distinguish between a data
product and its components, the following conventions will be used:
a simple object is an aggregation of data elements
a unit is an aggregation of objects
a data product is the highest level aggregation of units and/or simple objects
Data elements are collected into a format by the producer of the product. The names and
semantic descriptions of the elements are incorporated into a Data Element Dictionary
(DED). The format of the data elements is described using a Data Description Language
(DDL) or a Data Interchange Language (DIL), within a Data Description Record (DDR).
These are collected into a Description Data Unit (DDU) which is furnished to the data
