Dear Fedora Commons community,
We are currently studying the best approach for an institutional
repository using fedora, but are finding some difficulties.
It's easy to find fedora case studies on the Web, but our scenario is
somewhat different from all the others we are finding, even though we
find it hard to believe we are the firsts with such a use case.
Let me explain. We are trying to address two main concerns:
- Nowadays, all of our new documents are digital ones --- even if
they are paper documents, they are turned into a digital form when
they enter our institution. All of these documents are currently
stored on a very archaic repository, which doesn't provide us with the
control access we would like, doesn't properly handle unique object
identification, and doesn't really scale to a much larger number of
- The long term preservation of these objects hasn't really been
thought out before, but we want to start taking it into account.
So, our thoughts were to start using fedora, and ingesting digital
objects from the moment they appear on the organization. Our document
management system (which handles the document workflow) would make use
of the underlying fedora infrastructure by maintaining references to
the appropriate digital objects. Right on the moment our digital
objects were "born" they wouldn't have much associated metadata, but
it would grow as the object would be further used, throughout the
At the same time, taking our long term preservation concerns into
account, we have been looking at the OAIS model and at this repository
certification checklist: http://bibpurl.oclc.org/web/16712
These sound to us as very wise advises but, at the same time, poses some doubts:
- For certification purposes, the repository would have to maintain
*only* long-term-preservation objects, but if we ingest "newborn"
objects, that will not be the case. Only later the objects will be
evaluated and considered worthy of long-term preservation, or not ---
in which case they can be discarded.
- For our day to day administrative activities it makes perfect sense
to use a repository since the moment the document is created... does
this mean we will have to have a second repository to which to copy
the digital objects when the time comes? (this seems silly to us... Is
If someone has faced a similar scenario, we would really love to hear
your take on this matter. It seems to us that repository models are
being thought out mainly for end-of-the-line archiving (that is, when
they have served their main administrative purposes), but we think it
could be very useful to use them sooner on the document life-cycle.
Thank you in advance!