[Koha] Integrating digital repositories

[David Cook]

Hi Clint:

It looks like you've done a lot of work! I can't speak to Omeka, but I work
with DSpace and it's not too hard once you get used to it. It can be tough
to get support in Oceania as most of the developers seem to be in the US and
Europe, but they're pretty good with their listserv. I've played a bit with
Islandora, but haven't had the time to really go through it. It seems rather
massive though as it leverages the Fedora digital repository framework
(which other digital repository projects like Hydra use as well).

I originally worked on Bug 10662 in mid-2013, but I've had other priorities
come up over time. I've been thinking about reviving the code, but haven't
had the time and I'm not 100% sure how I want to structure it yet. Perhaps I
should write up a RFC at some point.

It sounds like Omeka's OAI-PMH plugin outputs MARC records, so that's good,
although I would be somewhat suspicious of the quality. I was transforming
DSpace records from OAI Dublin Core into MARC (as Koha only understands
MARC), and I was frequently disappointed with the results. This is partially
because MARC is a more "richer" metadata standard than Dublin Core. It was
also because OAI Dublin Core is way too basic. I found working with
Qualified Dublin Core to be at least a bit better, as it qualified dates and
identifiers way better. Ultimately, I decided that it would probably be
better to transform the records in DSpace and output them as MARC rather
than trying to do the conversion on the Koha side. Although I still wanted
to transform the records a bit so I could store the oai-identifier in the
MARC record (especially as some OAI-PMH repositories have rules about how
you can use their records - typically in the form of requiring some type of
provenance to be displayed in the record).

In any case, I'm still pondering Bug 10662. The code needs to be re-written
to take advantage of Koha::Schema and because the way I created the
harvesting objects was sub-optimal. I've also been debating (with myself)
about the best way to import records into Koha. Whether that's by using the
existing import functionality or whether to just directly insert in to the
database (so that users can't "undo" an import). I think the feature could
be implemented in numerous ways, so I'm still trying to think of the best
way to do it (rather than implement it and then have to try to re-write it
after people have already started using it).

As for experiences... at the moment we link DSpace and Koha using custom
scripts. I would prefer to go the OAI route but just haven't had the
time/money to finish the work :).

David Cook
Systems Librarian
Prosentient Systems
72/330 Wattle St, Ultimo, NSW 2007

> -----Original Message-----
> Date: Mon, 15 Sep 2014 15:01:30 +1200
> From: "clint.deckard" <clint.deckard at frontiers.co.nz>
> To: Koha-List <koha at lists.katipo.co.nz>
> Subject: [Koha] Integrating digital repositories
> Message-ID: <5416568A.9030402 at frontiers.co.nz>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
> 
> G'day all,
> 
> Working with Koha in schools I have often thought how more useful Koha
> could be if there was an easy way to integrate student created media eg.
> videos, podcasts, sound files etc
> But uploading files to a Koha server isn't straight forward. Uploading
> to sharing sites like Mega.co.nz  ;) is one way around, then cataloguing
> like any other 856u item, however having the file locally would be better.
> Another use case I came across recently required a local digital
> repository and this has focused me on the possibilities.
> 
> Here is a quick summary of my personal experiences. I am not an expert,
> this is new territory for me. I welcome any feedback and accept any
> mistakes I have made.
> 
> DSpace is released under the BSD licence and has a comprehensive feature
> set but not so easy to install. Others report not so easy to integrate
> with Koha although I didn't actually get DSpace working to experiment
> with it. Has a focus as an institutional repository. Backed by HP and
> MIT library.
> 
> Greenstone is released under GNU GPL, is predominantly installed on
> Windows but does have Linux and Mac versions. Developed in New Zealand
> at Waikato University with UNESCO backing. Reasonably easy to install
> in Linux but although functionally capable it feels 'clunky'.
> Development seems to have all but stalled.
> 
> Kete comes from the Horowhenua Library Trust (like Koha) and integrates
> very well with Koha from what others have said. Again, I didn't actually
> get Kete running on Debian; very nearly, but when I heard a new version
> might be available real soon, I chucked it in. The ruby on rails part
> was the hard bit for me.
> Has a focus as a community driven repository.
> development has had a hiatus of a couple of years but, rumour has it, a
> new version is almost complete. I am optimisic for Kete's future and
> will check back soon on developments.
> 
> Omeka is also GPL, actively developed and lots of functionality with the
> use of plugins. Very easy to install and easy to use. Shows a lot of
> promise.
> I added the OAI-PMH repository plugin to Omeka and then harvested
> records via the MarcEdit harvester tool for import into Koha.
> 
> This Koha enhancment
> Bug 10662 - Build OAI-PMH Harvesting Client
> sounds really useful in linking the digital archive with Koha.
> 
> I would be pleased to hear others experiences with digital repositories.
> 
> Clint.
> 
> 
> 
> 
> --
> Clint Deckard
> 
> Anact Open Source Solutions
> RD1, Takapau, New Zealand
>