Recently during a discussion on a Whatsapp group of professional colleagues from the North East, a topic that came up – what was better suited for setting up an ETD repository at their academic institutions? As expected, most sided with DSpace, while a few suggested Eprints. I decided to introduce Koha ILS into the picture. For most, this was a rather surprising suggestion.
The ground reality
95% of Indian ETD operators do very little than scanning up a batch of printed document (bound thesis volumes) or born-digital electronic copy of the theses, make it into a PDF file, throw in some metadata together about the item and plug that into (usually) DSpace. The benefits of DSpace being statistics, organisation into collections and community (user groups), embargo capability, faceted and full text searches across the metadata. There is of course the other point of persistent URLs to the hosted resources via HANDLE system. But given that very few Indian repos actually invest in a Handle.net account, it is quite a moot point.
The argument for Koha
Utilizing MARC21, Koha already offers a highly granular and extensive metadata regime. Its capabilities include collections, search facets, full-text search on the metadata. And since Koha’s version 3.22, the capability to directly host files linked to the bibliographic records. Basically it offers everything that 95% of Indian institutions look for when they are planning to setup a repository. Compared to DSpace, LIS professionals are already better acquainted with Koha (or as many in India like to call it – “KOHA” :-P) as it is the de-facto open source LMS. Further, by using Koha for hosting the repository, the institutions and professionals who are already using Koha as their LMS, gain one key advantage; they no longer need to maintain two separate and very dissimilar systems that use completely separate software application stacks (LAMP vs Java; MySQL vs PostgreSQL etc).
Step 1 – building a MARC21 ETD Framework
I had promised my colleagues to setup a demo ETD using Koha so that they could try it out. After all, there’s no better proof of the pudding than eating it. The first challenge to that is the while Koha does ship with quite a few MARC21 frameworks, an ETD framework is not one of them.
As my starting point, I turned to ETD-MS v1.1, which is considered to be something of a gold standard for ETDs. Taking the help of this page, I worked out a single paged worksheet for cataloging ETDs on Koha. The result looked like this:
Step 2- Usability and maintainability
Of course this posed a slight problem, how will a cataloger accustomed to / trained on DSpace’s metadata namespace correctly do the crosswalk? To solve this in lines with Koha’s recommended Best Practices for UI mods, jQuery statements were included into Koha’s “intranetuserjs” and “intranetusercss” system preferences. The final outcome was this:
This approach has three (03) key benefits – (a) that you DO NOT need to touch Koha’s Template::Toolkit based templates (i.e. the .tt files), the changes made are stored in your database and applied during runtime; (b) these changes remain persistent and works across Koha’s monthly version upgrade cycle (since we didn’t change the .tt files) and (c) our other MARC21 templates are left alone, only the ETD framework is thus modified.
The jquery snippet is available on Github as a gist, as is the CSS includes into intranetusercss system preference. There is also a third gist which lists the 04 authorised value categories and their constituent sample options that help us introduce a consistent controlled vocabulary in our metadata.
The demo ETD is available at http://etddemo-opac.l2c2.co.in/. If you wish to access the staff client back-end, contact me via the comments section of this post.