Documents

Documents

Overview

Representation of data and metadata for radio science applications are of fundamental importance. Clean object definitions, interfaces, and formats help ensure that instruments are easy to design and implement. Good metadata helps to make data understandable over the long term, enable debugging, and helps to avoid mis-interpretation of data.

Our approach is to define small independent objects each of which solves the representation requirements of a limited domain. These objects are then combined using a Namespace to provide solutions for more complex representation problems. When common patterns occur across different data objects we attempt to identify them, consider possible solutions, and understand the implications of them in real implementations. In some cases this leads to standardized approaches for solving the problem and associated software implementations.

In general for persistence and serialization we use [WWW] HDF5 as the format of choice. This format is a widely used standard, is open source, and well supported on a wide range of computing platforms. The HDF5 API allows representation of the namespace within a single file, is self describing as to content, and is high performance for large data sets. Built in compression and serialization support also enable the uniform and systematic use of objects represented in this manner as well as their transmission over data networks.

Patterns

Object Definitions

Standards

Radio Science Use Cases

Legacy

last edited 2009-01-23 19:00:47 by FrankLind