| Quick-Start Introduction to PDS Archiving |
 |
What We Need from Data Reviewers
The purpose of the review process is to ensure that the data are complete,
intelligible, interpretable, and of archival quality. All these
attributes are subjective to some extent, and reviewers should always bear
in mind that PDS is intended to be a research archive format, not a
working data set format.
Notwithstanding, following are some guidelines for evaluation.
Completeness
Completeness means that the data set contains all the documentation, calibration
and ancillary information needed to interpret the data. Any files used as
input to create the data files should be included (for example, ephemerides
for the spacecraft or target body). Generally speaking, for archive purposes
we prefer to have primitive data (which has perhaps been mechanically reduced
but has not yet had any scientific interpretation applied in the process),
as well as the final, interpreted data set (e.g., the project data set for a
spacecraft mission).
Completeness does not necessarily mean that every single bit produced by the
project has been saved. If the data supplier or the review panel determine
that some section of the data are scientifically meaningless (due to instrument
failure, for example), those data can and should
be deleted from the data set prior to
archiving.
Intelligibility
Intelligibility is concerned with the physical and logical formatting of the
data files.
Electronic data are considered intelligible if they are readable by any
reasonable computer system operated by a knowledgeable user who has access
to the data set documentation.
We assume that such a user and system can display any ASCII files included in
the data set without difficulty, and thus has access to the documentation
included with the data set.
The documentation should provide sufficient details about the physical and
logical format for the user to either read the data directly into an
application, or reformat the data as necessary to use it.
Intelligibility does not mean that the data must be in the most
convenient format. Rather, considerations of long-term archiving and
support for future research play the
major role in this determination. In other words, given the choice between
a format which is convenient but specific to some software application(s)
and a format which is based on a long-lived albeit inconvenient standard, the
standard format should be given preference. For ASCII data, formats which
make the data easy to read visually are preferred, since visual inspection is
a tried-and-true method for reading data and is unlikely to be
supplanted.
Interpretability
Whether the data are interpretable or not depends largely on the documentation.
At the record level, the documentation should define
the meaning, units of measure and
significance of each field.
At higher levels the documentation should describe how the data were
collected and reduced to their current state.
Calibration files and other ancillary data should be available and
themselves documented.
Interpretability should be measured with respect to a knowledgeable
professional. That is, interpretability does not mean that the documentation
must explain the data and their significance at the level of the general lay
public.
Archival Quality
Fitness for archiving should be interpreted with respect to the
research archive goals of the PDS.
It is neither the intention nor the purpose of the PDS to make immediately
available the latest revision of every possible data set; nor is it
desirable to congest the archive with data of questionable quality.
The opinions of the outside reviewers are weighed very heavily in
determining whether a data set is of archive quality.
|
