OARS: Open Access Repository System

On Friday 19th September 2008 we signed-off the development phase of the OARS Project after the sixth and final development cycle. At the beginning of this cycle we agreed that the basic deliverables were:

• Make it possible to re-ingest PDF, text, and thumbnails when editing metadata
• Re-sort by relevance, title and date in search results, by clicking on a heading on this page
• Enhance search engine with regard to foreign language text and search terms
• Allow items to be ‘hidden’ so that they don’t show up in User Interface searches

The final functionality was delivered by Aptivate on Monday 8th September 2008. The only functionality that was not factored in by agreement was ‘re-sort by relevance, title and date in search results’.

During the 2 weeks that followed, the application and its data was loaded into its final locations for live running - i.e. all repository data and its search engine hosted at the Oxford University Research Archive (ORA), and the ‘front end’ application hosted here at Forced Migration Online (FMO). Once that was completed, final acceptance testing was conducted by the FMO team, which successfully led to the sign-off.

The remaining tasks are to customise the application to add help pages, create styling for all web pages, and to configure url addresses etc to comply with the existing FMO web site. This work is being undertaken by the FMO team and is due to finish at the end of November 2008, at which point the new OARS repository system will become a live component of the FMO web site.

The current issue of the journal of the Association of Librarians and Information professionals in the Social Sciences (ALISS) includes an article about the OARS project. The July 2008 issue (vol. 3, no. 4) of ALISS Quarterly focuses on digital repositories. The article by the FMO team broadly outlines the main objectives and activities of the OARS project, why we felt it important to replace the current repository system and the technical challenges and solutions adopted to achieve this implementation. The article is available in full text (PDF file).

On Wednesday 23rd July we had the fifth meeting with our developers, which happens at the end of each development cycle (typically lasting 2 weeks). At the beginning of this cycle we agreed that the basic deliverables were:

• Add language as a filter on advanced search
• Show ‘Document Type’, ‘No. of pages’ and ‘Sourced From’ (formerly ‘Collection’) on search results page
• Show thumbnail, ‘No. of pages’ and ‘File Size’ on document / journal ‘landing page’ (i.e. metadata detail page), and facilitate ingestion of thumbnail along with document / journal
• Add css classes to web pages to facilitate page desgin

The functionality was delivered on Wednesday 23rd July. There are some outstanding issues to be resolved for this development cycle, and these will be resolved in the sixth and final cycle. Once the development and initial bug fixes are complete for the final cycle we will move the data (i.e. Fedora Repository) and the user/management interface to their live locations before embarking on final acceptance testing.

The priorities and deliverables for the final cycle were agreed, and briefly are:

• Make it possible to re-ingest PDF, text, and thumbnail when editing metadata
• Re-sort by relevance, title and date in search results, by clicking on a heading on this page
• Enhance search engine with regard to foreign language text and search terms
• Allow items to be “hidden” so that they don’t show up in User Interface searches

On Wednesday 2nd July we had the fourth meeting with our developers, which happens at the end of each development cycle (typically lasting 2 weeks). At the beginning of this cycle we agreed that the basic deliverables were:

• Ingest and index full text and all metadata from a new PDF document
• Add metadata field search - Title, Author, Collection etc
• Enable full index of all articles in journal issues for searching
• Enable editing of all metadata items in Management Interface, and allow certain metadata items to have multiple entries

The functionality was delivered on Wednesday 2nd July. There are some outstanding issues to be resolved for this development cycle, and these will be resolved over the next cycles. This meeting was combined with user evaluation sessions devised to solicit feedback about the way in which the current repository interface is used and to refine the functionality required of the new interface.

There were 7 participants, ranging from novice to expert users of FMO. Each session enabled one participant to use the current interface and comment on its strengths and weaknesses. The feedback from these sessions was very interesting. As we supposed, some of the more esoteric functionality of the current system is not used/understood, and what seems to be required is pretty much what we are developing - a more open system, with greater searchability via search engines such as google, better access from subject portals, and a slightly simpler and more intuitive user interface. Several recommendations for improved functionality were expressed and will be factored into the ongoing development.

The next development cycle will be followed by a development meeting on Wednesday 23rd July. The priorities and deliverables for this next cycle were agreed, and briefly are:

• Add language as a filter on advanced search
• Show ‘Document Type’, ‘No. of pages’ and ‘Sourced From’ (formerly ‘Collection’) on search results page
• Show thumbnail, ‘No. of pages’ and ‘File Size’ on document / journal ‘landing page’ (i.e. metadata detail page), and facilitate ingestion of thumbnail along with document / journal
• Add css classes to web pages to facilitate page desgin

On Tuesday 17th June we had the third meeting with our developers, which happens at the end of each development cycle (typically lasting 2 weeks). At the beginning of this cycle we agreed that the basic deliverables were:

• Index full text and metadata of stored items for search engines
• Improved functionality with regard to search results lists
• New functionality to allow view and download of full text for a document
• New functionality for inserting journal issues into the repository and indexing their full text

The functionality was delivered on Monday 16th June. There are some outstanding issues to be resolved for this development cycle, and these will be resolved over the next cycles. The next development cycle will be followed by a development meeting on Wednesday 2nd July. The priorities and deliverables for this next cycle were agreed, and briefly are:

• Ingest and index full text and all metadata from a new PDF document
• Add metadata field search - Title, Author, Collection etc
• Enable full index of all articles in journal issues for searching
• Enable editing of all metadata items in Management Interface, and allow certain metadata items to have multiple entries

On Monday 2nd June we had our second meeting with our developers, which happens at the end of each development cycle (typically lasting 2 weeks). At the beginning of this cycle we agreed that the basic deliverables were:

• Improved functionality around searching for documents, including use of search results lists
• Improvements to the management interface regarding metadata add/edit
• New functionality for inserting documents into the repository

The functionality was delivered on Friday 30th May and most was signed-off by the OARS team at the meeting. There are some outstanding issues to be resolved for the second development cycle, and this will be included in the next cycle of development, due to be completed on Monday 16th June. This will be followed by a development meeting on Tuesday 17th June. The priorities and deliverables for this next cycle were agreed, and briefly are:

• Index full text and metadata of stored items for search engines
• Improved functionality with regard to search results lists
• New functionality to allow view and download of full text for a document
• New functionality for inserting journal issues into the repository and indexing their full text

On Wednesday 14th May we had a meeting with our developers, which happens at the end of each development cycle (typically lasting 2 weeks). At the beginning of this cycle we agreed that basic functionality would be delivered around 2 scenarios covering the web-based user and management interfaces, and the underlying repository technology. These were:

• Finding and printing a known document
• Editing document metadata

The functionality was delivered on Tuesday 14th May and was signed-off by the OARS team at the meeting. The work done so far now provides a base for the next cycle of development, due to be completed on Friday 30th May, followed by a development meeting on Monday 2nd June. The priorities and deliverables for this next cycle were agreed, and briefly are:

• Improved functionality around searching for documents, including use of search results lists
• Improvements to the management interface regarding metadata add/edit
• New functionality for inserting documents into the repository

On Monday 28th April the software development part of the OARS project started. On Monday and Tuesday 28/29 April we had ‘Kick-off’ meetings with developers from our consultants (Aptivate). The two days were extremely productive, during which we discussed the schedule for for deliverables and testing, access to computing facilities and access to the repository at Oxford University Research Archive.

Basically increasing functionality will be delivered to the OARS Team on roughly a 2-weekly cycle over the next 12 weeks. The OARS team will test the delivery, advise on functionality issues and problems found, then set the priorities / deliverables for the following 2 weeks, and so on.

There will be an evaluation of the development process itself, with a view to highlighting the challenges we face and hopefully to provide lessons for future projects. This is a very important part of this phase of the project and will be managed by our evaluation consultant Jan Rae.

We are now finalising the repository development plans with our chosen developers Aptivate (http://www.aptivate.org), and work will commence in March. Aptivate will be working closely with our colleagues at the Oxford University Research Archive (http://ora.ouls.ox.ac.uk/access/) who will host the Fedora infrastructure for our new repository.

In December 2007 we wrote to all university departments and centres in the UK that conduct research, teaching and/or dissemination activities in issues related to migration, conflict, development and human rights, and they were asked to complete a questionnaire as a way to kick-off this consultation process. Questionnaires have now been returned and are being analysed.

Towards the end of November 2007 we interviewed for the post of OARS Digital Repository Software Developer. We had advertised the post widely - Oxford email lists, jobs.ac.uk, Fedora lists, slash dot, 43folders etc but received only 5 applications, of which 2 were invited for interview. One candidate subsequently declined and the other proved to be unsuitable for the post.

We are now in discussions with a team of developers at Aptivate (http://www.aptivate.org), an IT development company that has emerged from Cambridge University and specialises in IT development for ‘developing world’ projects. We have had connections with them for many years. We are hoping that that they will be able to undertake all of the work we had planned for our developer.