84
user. This methodology allows a complete understanding of the content of each simple
object.
Building a Data Product
The following diagram is an example of what a simple data product might look like. In this
case, the product contains information on the production of the product and an object of
identification information containing catalog data pertaining to the product (e.g. platform
ID, data type, time span). Following this are two Description Data Units (DDU), one for
the application data object and the other for the supplementary data object. Finally, there
are the instance data comprising an supplementary (or ancillary) data instance and the
instance of the application data itself (e.g. the image).
PHYSICAL PRODUCT INFORMATION
IDENTIFICATION INFORMATION
DDR FOR SUPPLEMENTARY DATA
DDR FOR APPLICATION DATA
SUPPLEMENTARY DATA INSTANCE
APPLICATION DATA INSTANCE
The exact content and order of the content of a product varies with the problem at hand.
The amounts and types of metadata and supplementary data may be more or less than that
shown. In other words, all of the information needed to understand and analyze some
collection of data should be packaged with it in some manner.
Most data products contain multiple instances of data and metadata. Each instance is a
group within the product containing, for example, an application data instance (e.g. an
image), a supplementary data instance (e.g. calibration data), and the identification of the
unit (e.g. time and location of the data). An example of such an instance is shown below.
This unit is termed an information unit.
IDENTIFICATION INFORMATION
SUPPLEMENTARY DATA INSTANCE
APPLICATION DATA INSTANCE
In this case, the overhead may be lessened by not repeating the data description with every
instance. Hence the product diagram looks like the following (& indicates that this unit may
be repeated many times):
