I think it's actually worth interrogating and getting specific about what
we mean by "preservation features".
I think they may not actually be all that complicated or hard to add on to
nearly any solution. I think an actual 'repository solution' may actually
not be as complicated as people assume when you actually specify it.
The main preservation feature people actually use, "fixity", is just taking
a checksum of a file (perhaps using SHA1), storing it somewhere, and then
later checking to make sure the file still has the same checksum, and
alerting if it does not. This is a relatively simple feature to add to any
If we look at what Fedora, for instance, actually advertises as it's
preservation features, here:
Persistence: Okay, any system provides some kind of persistent storage,
and "Deposited files are stored on the filesystem in a predictable
location" seems to matter for backup purposes. I'm not sure this
particular implementation is actually a requirement for proper
back-up-ableness, but if it is it's not too hard to do.
Fixity and Audit: See above. Pretty straightforward to implement in any
Versioning: This is the most challenging feature to implement, but as a
consequence our actually existing systems often implement it incompletely
if at all, and this does not seem to bother many. While Fedora can do
versioning for metadata or attached files, Sufia/Hyrax only provide UI/API
for versioning attached files, and the functionality is pretty limited at
that. I think most people using Fedora through sufia or hyrax are not
making use of versioning much if at all. The experimental "Valkyrie"
system (a new persistence layer for Samvera (nee hydra)) does not yet
implement versioning of any kind, and I believe the institutions planning
on adopting Valkyrie do not require it for an initial production
Import/export: Of course some kind of import/export is trivial, but this
point also suggests an _interoperable_ import/export. "Integration with
preservation systems external to Fedora", "Avoidance of 'platform lock-in'"
-- while Fedora aims to do this via an import/export in RDF, it turns out
"it's in RDF" isn't enough to provide interoperability, and I'm not sure
interoperable imports/exports actually exist at all to this day
realistically. It turns out to be a very hard problem, more about
modelling, and the difficulty of consistent modelling that is expressive
enough for any system, than about technical implementation.
I think our community over-mystifies "preservation features" to some
extent. The ones we actually have are all either simple enough that it is
not hard to implement in any system, or complicated enough that actually
existing systems don't even do a great job of them, and that doesn't seem
to be stopping anyone.
I think there's no fundamental reason much simpler preservation solutions
can't be available, than what we generally have access to now. Simpler,
with much lower total-cost-of-ownership. Especially for smaller and simpler
institutions. And think it's a great idea for people to experiment in that
On Wed, Oct 25, 2017 at 8:59 PM, Bryan Brown <[log in to unmask]> wrote:
> Theres nothing wrong with what you are describing if its all your
> institution needs, but I would be careful about promoting that as an IR. An
> IR is much more than a bunch of documents. The metadata modelling,
> preservation features and indexing that you want to leave out are what
> makes it a repository. Also, the infrastructure you are describing may lack
> flexibility in the future if you decide you want to add new features to it.
> Bryan J. Brown
> Repository Developer
> Technology & Digital Scholarship Division
> Florida State University Libraries
> From: Code for Libraries <[log in to unmask]> on behalf of Josh
> Welker <[log in to unmask]>
> Sent: Wednesday, October 25, 2017 2:51:34 PM
> To: [log in to unmask]
> Subject: [CODE4LIB] Lightweight IR infrastructure
> We're a mid-sized university library (10,000 fte) trying to get an IR off
> the ground to showcase student and faculty research. We've had a DSpace
> instance running for several years, but we use so few of its features that
> DSpace ends up being more trouble than it is worth. In particular, it's
> very frustrating to deal with metadata editing, file management, the Handle
> URL system, and HTML/CSS theming.
> I am considering leaving the DSpace model in favor of our "IR" just being a
> glorified FTP site that MARC records in our catalog can point to. I might
> even build a tiny frontend using some scripting language to add IP
> authentication, URL redirect stuff, or a Google Scholar interface, but
> that's really it. No metadata modelling, no preservation features, no
> Does anyone have experience using a very small, file-based (as opposed to
> database-driven) application as a foundation for an IR? Are there any
> problems I should anticipate?
> Joshua Welker
> Information Technology Librarian
> James C. Kirkpatrick Library
> University of Central Missouri
> Warrensburg, MO 64093
> JCKL 2260