Why file formats?
File formats may not seem like part of visualization, but they are an integral part
of the cycle from the user to the application to the visualization and back. There
may be a format preferred for your problem domain, but that is often not the case.
The most common practice is to use text files with annotation information, like
the date and run parameters, encoded in the filename. With the help of zip compression,
this can be a moderately successful approach to archiving. The longevity of your
data, and even the short-term accuracy, can depend strongly on your ability to record
where data came from, what model it measures. To that end, the favored scientific
data formats contain, within a single binary file, whole directories of information,
from text comments and measurement units to derived images and movies. Shown below
are some of the gold standards.
Why not ASCII?
- Larger file size
- ten digits of eight bytes or twelve bytes.
- Visualization breeds lots of files
- Slower to read and write
- Lonely little numbers - ASCII numbers have no units, no mention
of upper and lower bounds for checking errors.
