I agree: there is no necessary connection between an archival package
and a represenatation. There might be multiple representations in an
archival package and vice versa. And there are multiple ways of
construction and transfer.
In our use cases we want to record that information, because the naming
of an Information Package has a defined meaning and is assumed to be the
representation. But names will change after transmission between two
And I agree there might be much more complex scenarios. But: is there a
need for a more complex relational model in that case?
In our case it is simple because it is predefined. I would be glad to
have the opportunity to record the former name of the representation (in
our case archival package), because there is no way to do this in PREMIS
Evan Owens schrieb:
> A fascinating question! Is archival package the same thing as
> representation? In all cases?
> I understand the value of original file name and even knowing
> information about how the files were packaged together for the purposes
> of transmission to the archive, but it is not obvious to me that that
> has anything to do with representation...at least not in all cases. We
> get large tar files with many articles each consisting of many files;
> the name of the original tar file is interesting and worth preserving
> somewhere, but it is not a property of a "representation" as the
> transmission package in this case is of many works or intellectual
> objects. I could imagine the opposite as well: receiving a large
> complex intellectual work in installments consisting of several
> We do get content packaged one tar or zip per article. In that case, it
> would seem to appear as if there is an originalName for a given
> representation. But I would argue that that is a coincidence and not a
> universal truth.
> There are even more complex examples that are worth considering: the
> archival package may include multiple representations of the same
> intellectual object where in the worst case some files belong to more
> than one representation.
> It seems to me that it is not safe to assume a simple one to one
> mappings of how things are packaged for transmission and how they are
> preserved. In practice this can easily become a complicated one to many
> problem. That would suggest that it needs to be stored in it is own
> data structure, one that can describe such relationships.
> If it is really important to be able to reconstruct exactly how objects
> were received it might be easiest just to preserve the exact
> transmission intact.
> Evan Owens, Chief Technology Officer
> Portico www.portico.org
> [log in to unmask] (609) 258 8230
> 228 Alexander Street, Princeton NJ 08550
> -----Original Message-----
> From: PREMIS Implementors Group Forum [mailto:[log in to unmask]] On Behalf Of
> Sent: Tuesday, October 24, 2006 11:43 AM
> To: [log in to unmask]
> Subject: [PIG] originalName in representation
> Hi list,
> I have to bother you again. I found another problematic Semantic unit:
> It is not applicable to a representation. But if one comes to the use
> case where you would like to exchange Information Packages in a file it
> would be advatageous to have the original name of an archival package
> which is exchanged between different systems over the network.
> Definition The name of the object as submitted to or harvested by the
> repository, before any renaming by the repository.
> Rationale The name used within the preservation repository may not be
> known outside of the repository. A depositor might need to request a
> file by its original name. Also, the repository may need to reconstruct
> internal links for dissemination.
> Data constraint None
> Object category Representation File Bitstream Applicability Not
> applicable Applicable Not applicable
> Do others also see this problem?