Reflections from OR2010: Part 2

Another activity I was involved in at OR2010 was the Developer Challenge as a ‘non-techie’ judge. Organised by the DevSci project (managed by UKOLN, funded by JISC), this year’s challenge was to ‘create a functioning repository user-interface, presenting a single metadata record which includes as many automatically created, useful links to related external content as possible.’

The winning entry was from Richard Davis and Rory McNicholl, University of London Computer Centre, who enhanced the records of the Linnean Collections, held on EPrints, which the ULCC are responsible for. As many of the metadata fields in the record as possible linked out to external sites- some general such as Google and Wikipedia and some more subject specific such as horticultural indexes. Although only one metadata record was demonstrated, the links which appeared in the record were determined by the entries on a master sheet (an excel spreadsheet) and therefore, would apply to all records within the repository. This development came out top as although the links weren’t truly automated, they were managed externally, it was felt that this was actually advantage for a non-techie repository manager could update for themselves rather than calling on the support of a tame developer.

Coming in a narrow second was ChEsis, presented by Sam Adams, University of Cambridge, which created links to enhance Chemistry e-thesis records. Links were available to show chemical structures of molecules/ crystals used/ created along with their mass spectrums. Fun links were also included such as the Last FM playlist of the student when their thesis was submitted and the BBC headlines for the day.

Another entry utilised OpenCalais to automatically create links from their created repository record. OpenCalais is a free to use service which automatically creates links from open content to other open content sites such as Twitter, YouTube, Flickr etc. It can be used to add a bit of fun to any open source web content such as a blog but be warned the links are automatic and you can’t necessarily restrict what content it links to!

A full write-up and videos of all the Developer Challenge entries is available via DevSci blog.

DSpace Add-Ons

The following are descriptions of a range of DSpace add-ons available to install which provide additional functionality to the software.

Commenting
The commenting feature brings informal communication capabilities to the DSpace environment. A threaded forum, or comments stream, can be attached to any web-page, community, collection, submitted item or e-person within DSpace. The add-on allows comments to be inserted by both anonymous users and authenticated ones while functionality for reviewing/moderating comments is also provided.

An example (in Portuguese) of comments appearing below a collection.

Compatible with DSpace 1.1 and 1.2 possibility of updating for DSpace 1.5.x


Controlled Vocabulary/Ontology
This add-on applies a subject classification system of the institutions choice to their DSpace instance. Once implemented the user chooses from the predefined taxonomy of keywords to describe items of information that are being submitted to the repository and that same taxonomy is used to find and access items held in the repository.

An example (in Portuguese) of a subject classification system in use.

Compatible with DSpace 1.1 and 1.2 possibility of updating for DSpace 1.5.x


Dublin Core Meta Toolkit
The Dublin Core Meta Toolkit gives DSpace administrators the ability to convert large amounts of information from their desktop database programs into DSpace compatible Dublin Core metadata. The toolkit provides a number of out-of-the-box database structures to ease data collection as well as enabling users to create custom converters for existing databases. The Toolkit is ideal for converting formats from Microsoft Access, MySQL and comma delimited value (CSV).

Compatible with DSpace 1.5.1


Embargo
Content submitted to a repository may be restricted by laws, policies, or contractual obligations that require the submitter not to publish or enable public access to the content for a period of time.This add-on allows DSpace administrators to build in functionality to handle embargoed items in the workflow. It allows for the metadata of the embargoed item to be indexed and viewed, but the full text of the item cannot be retrieved while the embargo is in force.

Compatible with DSpace 1.4.x, 1.5.x and 1.6


Format validator and virus check
This add-on provides rough-and-ready format checking by identifying that the file/bitstream extension matches formats verifiable by JHOVE. Currently DSpace accepts a deposit's file extension as gospel, so a user could tack a .txt extension onto a GIF and DSpace would assign the incorrect format to the file based on that incorrect extension. It also checks the file for the presence of viruses.

Compatible with DSpace 1.4.x and 1.5.x


Recommendation
This DSpace content-based recommendation feature automatically shows links to articles within the repository that a user is likely to be interested in by mapping items related to the document currently being visualized by the user. Similar to functionality seen on Amazon this feature can greatly improve user experience.

Compatible with DSpace 1.1 and 1.2 possibility of updating for DSpace 1.5.x


Request Copy
This add-on creates a semi-automated mechanism whereby would-be users can request and authors can email an individual copy of a full-text deposited within the repository whose full-text access privileges are set to restricted.

The purpose of this feature is to increase both the content deposited in an IR and its immediate usability by providing a way to accommodate the (frequently unfounded) worries of authors and their institutions about copyright infringement during any publisher embargo periods on public self-archiving.

The link is provided on all non-OA items and activates a form where the user requester must enter his/her email address and name, and may add a comment, and press a 'Request-a-copy' button. An email is sent to the depositor and the email message contains a token. Using that token, the author may reply, by just clicking in one of the two buttons available: 'Send Copy', 'Don't send copy'.

Compatible with DSpace 1.4.x and 1.5.x


Semantic Search for DSpace
Semantic Search allows intelligent search of DSpace content, using Semantic Web technologies and performs knowledge discovery on DSpace metadata. Semantic Search uses the science of meaning in language, to produce highly relevant search results.

Compatible with DSpace 1.4.2 possibility of updating for DSpace 1.5.x


Statistics
This add-on allows gathering, processing and presenting usage, content and administrative statistics from the repository. The system is based on components that can easily be configured, changed or extended, to respond to different information needs.

Compatible with DSpace 1.4.x and 1.5.x (JSPUI)


Tombstoning
The add-on allows a tombstone to be added when an item is withdrawn from the repository. The user selects from 3 reasons for withdrawing the item: 1. Removed from view by legal order; 2. Removed from view by the [authority doing removal]; 3. Removed from view at request of the author.

Compatible with DSpace 1.4.x and 1.5.x


Further details about this range of options are available here. Any WRN partner interested in discussing or investigating any of these DSpace add-ons should contact the team using the usual address wrnstaff@aber.ac.uk

Citation Count APIs

A couple of weeks ago Stuart Lewis disseminated to the repository community some work he had been doing on a citation count API for use in repositories. The API created by Stuart creates a plug-in to a repository item record showing the citation count for that item in the SCOPUS bibliographic index. The API works by matching the DOI included in the repository item record with the DOI in the SCOPUS record for the same item. Discussion within the mailing lists however, highlighted that it would be fairly easy to extend the API to match with an item’s title/ author/ year etc. if DOIs are not consistently present within repository item records. A blog post about the API and how to implement it within a repository can be found within Stuart’s blog. Further examples of how the citation count API can be used are available from the SCOPUS website.

An example of the SCOPUS citation count API in action within a repository record can be seen from The HKU (University of Hong Kong) Scholars Hub . This item record example also has citation counts for the item within Web of Science. Through some Google searching I hunted down the following blog post by Jonathan Rochkind at John Hopkins University highlighting the application of the “Link Article Match Retrieval Service” (described at the bottom of the webpage). If your institution has a subscription to WoS, this API will provide you with the citation counts for an item matched by DOI/ author/ title for use within a repository record.

The use of citation counts within repository item records can be extremely useful within the forthcoming REF as a way of demonstrating the level of ‘Impact’ an item has had. It can also be a way of enabling institutions to select the publications they wish to put forward for the REF, highlighting a well received piece of research.

The APIs described above will give you the total citation counts for the item within those specific bibliographic databases. For some institutional reporting the citation counts per year are necessary but specific subscriptions to WoS and SCOPUS services are necessary to access these.


If you would like more information about the use and implementation of citation count APIs within your repository please contact the WRN Team at wrnstaff@aber.ac.uk.

Text Mining for Scholarly Communications and Repositories Joint Workshop

The Text Mining workshop was held in Manchester on 28-29th October 2009 and focused on the challenges and priorities associated with integrating text mining technologies in applications to support scholarly communication and repository initiatives. With the vast amounts of information now available on the internet, the benefits provided by text mining for discovering relevant documents have become increasingly significant.

Professor Tony Hey presented the keynote and he spoke about the need for more intelligent data discovery in a multi-disciplinary and collaborative way for Science to move from data and information towards knowledge (DIKW). These complex technologies have been applied successfully to the Science domains, particularly chemistry and medicine and are being adopted by BioMedCentral and Elsevier. Rafael Sidi from Elsevier again spoke of information overload and the importance of building applications on top of content using open Application Programming Interfaces (APIs). This would allow interoperability and collaboration between publishers' collections and the potential for free access to content with subscriptions for the added services.

Emma Tonkin's overview of the FixRep Project at UKOLN that is examining text mining techniques for automated metadata extraction was particularly relevant to the repository world. Presentations are now available online at The National Centre for Text Mining.

JISC Deposit Show & Tell

I attended the JISC Deposit Show & Tell event held at Birkbeck, London on Monday 12th October. The aim of the event was to identify deposit tools or combinations of tools that would clearly benefit repository users and to plot a path for those tools toward widespread and sustainable take-up. JISC's funding roadmap includes provision for sustained improvements to the 'deposit' process and it is hoped that the outcomes from the event will inform JISC's planning.

The first half of the day provided a stage for developers to 'show & tell' the deposit tools they have been working on so that a list of features/functions that have been used in a real end user deposit processes could be created. The second half of the day was spent mock prototyping projects that could further build and distribute the next generation of deposit tools to specific end users.

Stuart Lewis from Auckland University presented an interesting GenericDeposit via email (and SWORD). The aim is to provide academics and researchers with a familiar interface - an email is sent to the 'repository' with the title in the subject heading, an abstract in the body and files for deposit are attached. The author is assumed to be the sender and an email reply is then received on the status of the deposit. Very neat and very simple!

EMLoader, demonstrated by Fred Howel enables easier deposit of research papers through bulk upload of bibliographic metadata. The functionality again uses SWORD and connects two existing services: the Depot, a UK repository for researchers who do not have other provision, and PublicationsList.org, a web site for researchers to build a web page listing their publications.

The day was well attended and brought developers together from a range of projects in a productive session - JISC will issue a call for funding based on some of these ideas.

Open Journal Systems (OJS)

During the last WRN video meeting, partners expressed an interest in Open Journal Systems (OJS) which is being used by libraries for publishing and managing journals at low cost. It is open source, has a flexible design, can be installed locally and is easy to use and administer. Morag Watson from Edinburgh University gave a very positive review of this system at the recent Edinburgh Repository Fringe 2009. There will be a further blog post covering this event in the next few days.