background image
Relationships and Aggregations
n area that is not developed to any
great extent in this metadata set is that
of relationships between objects. I have
already discussed the difference between
digital or, rather, physical objects (i.e. a data
file stored on a medium) and conceptual
objects (i.e. a publication or archival record
as presented on a screen), which is not
necessarily one-to-one.This conclusion
entails more complexity and has its impli-
cations for defining metadata.
Relationships in the proposed metadata set
are now identified as a subset of Context
Information (one of the sub-categories of
Preservation Description Information).
They consist of two sub-categories:
`Manifestation' and `Intellectual Content'.
These relationships can refer either to other
manifestations of the same object (= the
same content) or, in the case of Intellectual
Content, to relationships between this and
other objects (with related content).As such,
the relationships seem to refer to physical
data objects.They cover only part of the
relationships necessary to represent the com-
plex network of physical and conceptual
objects. Since an intellectual object, either
a record or set of records or a publication,
can consist of several digital components
and vice versa, as discussed above, this area
needs to be enhanced.
n order to accommodate the identified
complexity it is necessary to distinguish
at least between the following categories
of relationships:
- relationships between intellectual
objects: for instance, components of one
collection or, in the case of records, the
components of a case file or an (archival)
fonds or a series (representing different
aggregation levels) and their position
within that aggregation. In the archival
context this is referred to as `documen-
tary context';
- relationships between the (structural)
components of one intellectual object,
e.g. the pages of a book, the elements
of a record, the components of a
organisation responsible. In this metadata
set this is represented by the sub-category
(of Context Information) `Reason for
creation', defined as `documents informa-
tion about why a content data object was
created'.The accompanying explanation
indicates that it concerns mainly why a
physical object was created. It limits the
scope of contextual information to `infor-
mational requirements associated with
managing the preservation process'. As
such, it only refers to the role a certain
data file plays, e.g. master file or similar,
a rather technical approach.
or information that helps to under-
stand the background of the digital
object, one is referred to `Representation
Information' (i.e. the section Content
Data Object Description). In this part the
element `Documentation' is meant to pro-
vide the documentation necessary to
interpret the `Content Data Object' and is
assumed to be a link to where the docu-
mentation is (e.g. a URL). Perhaps this
may represent the kind of information I
mentioned above. Apparently this infor-
mation is not considered as preservation
metadata. Nonetheless, it is essential for
understanding the intellectual objects
(publications, documents or records) as
represented on the screen and for being
able to identify their authenticity.
Furthermore, this information must be
inextricably linked to the (intellectual)
objects themselves. As such, it also has to
be preserved and may be an object in
itself. And in the case of records, this kind
of information or metadata will be created
electronically and must be preserved as
such. In the Information Model and the
derived preservation metadata set this part
is not adequately addressed. In order to be
able to preserve (archival) records it will
therefore be necessary to extend the informa-
tion model with another class of informa-
tion that refers to business context. Such a
subset could provide a structure for de-
scribing what in archival terminology is
called information about `provenance' (with
a different meaning from that in OAIS).
reliable. In other words, it should be possible
to position a record in the time and con-
text in which it purports to originate.
After all, if we know somebody or some-
thing (or think we know), we are able to
establish whether we can trust him, her or
it, or not. In order to enable this judgement
or assessment we need information that
can answer those questions. In a digital
environment, in particular, authenticity
has become an issue, because digital docu-
ments or records by their very nature are
intangible and volatile, and easy to tamper
with. Apart from information about their
origin, information about the management
of the records or digital objects is necessary
to be able to assess what happened since
they were captured and whether some-
thing may have occurred that has affected
them in a negative sense. Finally, the (con-
ceptual) object itself has to be described:
what are its essential characteristics? To
some extent these categories of informa-
tion may be found in the proposed meta-
data set.Thus it can be indicated, for
example, under `context information' why
an object has been created, be it in a rather
technical sense, and under `provenance
information' what `events' have taken
place (management history), while under
Content Data Object Description `signifi-
cant properties' are included, identifying
the characteristics of the object that should
be preserved.The question is, does this
really meet the requirements for main-
taining authenticity? It would have been
helpful if there had been more acknow-
ledgement of the issue of authenticity and
the requirements for it, and if the Working
Group had provided some background
information about its view and considera-
tions on this aspect and to what extent it
is included or not.
s indicated, context information in
an archival sense means information
about the context in which records are
created, i.e. the business activity and the