background image
DigiCULT
.
Info
16
information objects, have or may have a
relationship with each other, but in other
cases may not, or may overlap, or may
have different meanings.The term `digital
object' is the most ambiguous, since it
refers to both conceptual and technical
aspects, i.e. the intellectual content and its
form, and the way it is digitally represented.
At least, this is the way I will be using it
in this article.
I
n this respect it is essential to know the
underlying concepts in order to under-
stand the metadata set, and I think more
discussion is needed in this area. It is,
despite some definitions, not always quite
clear what these concepts are or what
they really mean.
Scope
T
he preservation metadata set, as pre-
sented, is a kind of compilation and
elaboration of the work on metadata
undertaken in different projects (NEDLIB
and CEDARS) and by different organisa-
tions (OCLC and NLA), and as such can
be considered very useful.Thus the back-
ground is mainly the library community,
as is also obvious looking at the composi-
tion of the Working Group.The title sug-
gests, however, that this proposed set of
metadata is applicable to all digital objects
in all domains. On p.3 it is indicated that
it should `accommodate the needs of the
library community, along with other
institutions tasked with the long-term
management of information in digital
form'.This is, however, as I will indicate
below, not the case and this is one of the
restrictions to this proposed set.
I
n line with the OAIS model, the point
of view taken is that of the custodian
not being the creator. It supposes that at
some moment in time digital objects have
to be transferred, brought in, harvested, or
ingested into the organisation and system
the custodian is managing. Prior to this
ingest, the objects are under the control of
other organisations, such as publishers or
record-creating organisations.The extent
to which the management of objects can
be influenced differs with respect to the
type of objects. In the case of (govern-
ment) records, legislation governs their
creation and management, whereas, in the
case of publications, the influence will be
mostly based on agreements between pro-
ducers, publishers and preservers.
T
his includes also the way metadata are
created and maintained. Ideally, there
should be continuous and consistent crea-
tion and management of the metadata that
accompany the digital objects. In practice,
there is still a lot to achieve in this respect.
So, although the suggestion may some-
times be otherwise, preservation metadata
do not only apply to what is under the
custody of a cultural or other preserving
institution, but should be applied to the
whole lifecycle of digital objects.This
implies lifecycle management. I use the
term lifecycle here as a chain of stages,
such as conception or design, creation,
maintenance and final disposition (as in
the case of government records, destruc-
tion or `sentenced' for having continuous
value). During that lifecycle the use of
digital (and other) objects may change
over time or across domains, because of
different contexts, and this will have an
impact on the creation and management
of metadata. Preservation can be viewed
as part of maintenance.
`P
reservation metadata' are seen as
information necessary to support
preservation processes. In the report, how-
ever, no link is made to these processes.
An analysis of the processes and the kind
of information needed to perform them
would have been very useful to better
understand this proposed metadata set.
Without such an analysis it seems more
difficult to identify the metadata needed.
The Working Group may have considered
it implicit since the set is based on the
OAIS information model, which again is
derived from the functional model.The
high level of that model, however, does
not make the required metadata immedia-
tely clear. In this respect it is also essential
to understand the nature of the digital
objects that are to be preserved.There is
an explicit statement that `no assumptions
about type or structure of digital resource'
are made and that no particular preserva-
tion strategies are taken into account (p.3).
But, by taking library community needs as
leading (albeit implicitly), the approach is
already restricting the types of digital
objects. Managing different types of `digital
objects', e.g. publications and records, may
require not entirely similar sets of metadata.
A
nother issue is that of the require-
ments governing the preservation
processes. Approaching this from the
object itself, as is the case in this report, is
not sufficient.There needs to be insight
and, as a consequence, also metadata about
the preservation strategies, policies and
methods, together with the context in
which the preservation takes place.
Although the OAIS model contains func-
tions like `preservation planning' and
`administration', these are not addressed in
the report, perhaps because they are not
(yet) reflected in the information model.
For a complete overview of preservation
metadata, nonetheless, all functions have
to be included.
Digital objects
A
crucial question in understanding
(long-term) preservation of digital
objects is: what is a digital object? As already
indicated, this seems not as clear-cut as
one would expect. In the OAIS model an
object of preservation is identified as a data
object that can be either physical or digital.
A digital (data) object is defined here as
`an object composed of a set of bit se-
quences', therefore it has to be seen as a
(data) file that is stored on the disk and
not as the `thing' that is represented on the
screen.The data object is accompanied by
so-called `representation information',
which documents the way the data object
has to be interpreted in order to present it
and to enable intended users to view and
understand it.Together, they form the
information object. It is a rather technical
approach that is taken here. Perhaps the
real question should be, what do we want