Monday, 25 October 2010

Lessons Learned

Procedural Improvements

In hindsight, the lack of opportunity to beta iFind Discover in a soft launch was critical as it meant that several major problems caused disruption to services. These problems only revealed themselves under heavy usage and included resource allocation issues with the Jangle software used to connect to Swansea Metropolitan Library, Open SSL and LDAP connectivity and system resource allocation. Identifying these problems at an earlier stage may also have enabled complete two missing elements to the catalogue; a stable SSL certificate and the automated update of records.

The feasibility of a Multi LMS Discovery Catalogue

Our work on VuFind has shown that it is technically possible to modify the core to allow it to operate in a multi-LMS environment. The degree to which system functionality can be replicated and improved on is very much dependant on the LMS involved and relies on the openness of data and system services.

The major issues for the implementation of Multi LMS Discovery Catalogue are system performance and the accessibility of data in the user interface. Searching more than one system has inevitable implications for search time and produces greater technical challenges such as negotiating firewalls and merging data sets. The implications of differences in ICT policy, data sharing and commitment amongst participating institutions also need to be taken into consideration. In addition, presenting multiple sets of data also requires careful planning or users will be swamped with information, unable to quickly identify the details which are important to them.

Beta feedback

The Beta version of iFind Discover went live towards the end of Summer 2010 and was intended to be functionally complete. Most of the consolation from this point resolved around the user interface and the information displayed. Training and introduction sessions were arranged all three institutions and users were encouraged to submit feedback directly via e-mail or through the inbuilt feedback form. To facilitate the collection and organisation of this feedback, a blog (http://ifind.swwhep.ac.uk/blog) was produced in order to list project news, user suggestions for new developments, new features from the core developers, new features in development by the project staff and confirmed developments. (For a summary of the feedback received – please see below)

Throughout the feedback and consultation process, very little feedback was received users from Swansea Metropolitan University and Trinity Saint David. As a result, there may be a bias in the structure and nature of information displayed in the catalogue but without further feedback or engagement in the development process, this is unavoidable.

iFind Discover Beta became the default discovery tool for Swansea University at the beginning of the 2010/2011 academic term. A soft launch with iFind Discover Beta as an optional catalogue would have been preferable as it would have allowed us to identify the system instabilities and bugs with less core service disruptions. It could then have become the default discovery interface for Swansea University at the beginning of 2011. Unfortunately, due to the time constraints of the SWWHEP project, this was untenable.

The Consultation process

The consolation process for the development of the new catalogue was due to begin at the beginning of last term academic term of 2009/2010 and continue for the duration summer. Unfortunately, due to operational difficulties in completing the modifications required for a multi-LMS environment, consultation actually began after the end of the last academic term of 2009; as a result, the number of students and staff involved in the consultation process was dramatically reduced and less over all time was available for examining data presentation and identifying system bugs and problems.

The loss of “bulk” feedback from students was particularly troublesome as it meant that issues with system reliability under heavy usage were not uncovered until much later in the process- this in turn led to periods of system down time and service disruption.

To help offset these difficulties, an unobtrusive feedback form was added to each page of the beta catalogue to collect information from users who were not part of the official consultation process.

An extremely useful opportunity to ascertain the potential “market” for a multi-lms discovery catalogue was afforded by a trip to the VuFind 2.0 conference in Philadelphia. Though there is more scope there for implementation of VuFind in consortia which share the same LMS, sessions on the user interface revealed provided extremely useful best practice studies which the project team was eventually able to use in addressing some of the initial feedback.

One of the flaws in the consultation process was that too much focus was given to highlighting and developing the new advanced features of VuFind. The result was that not enough consideration was given to identifying services which were part of the existing catalogues such as Voyager’s Book Bag feature, which were not part of VuFind.

Establishing requirements

Before any development or consolation could begin, it was necessary to identify where the core VuFind system would need to modified to achieve the goals set out in the original Virtual Academic Library project. These goals were:

“A single portal … to allow cross-searching of library catalogues and access to shared e-resources”

“an essential corollary of reciprocal access and borrowing”

After analysing the VuFind core, it became apparent that the project would require 1 Solr record for up to three institutions, connections and drivers to three Library Management Systems, three methods of authentication and the development of additional functions to allow users to place and cancel holds and renew items.

Thursday, 21 October 2010

The best laid plans of mice and men ....

The original timetable for this project looked like this:

March 2010 – Initial “go live” of public beta service

June 2010 – completion of feedback-gathering exercise from beta testing

August 2010 – launch of full, live service

Septemebr 2010 – project review

October 2010 – Options for sustainability of the service beyond the SWWHEP project funding period and handover plan

It didn’t quite work out that way!

There was an embryonic version of the beta service available by the end of March, but there was still too much work to be done on it before it could be properly publicized or used. The reasons for this were various, but included dependencies on outside sources for the development of addition Jangle features, and problems getting permission for access to one of the institution’s LMS’s in order to enable and develop its VuFind connector.

As a result, the service wasn’t fully in beta until the summer, when feedback (mainly from staff) started rolling in. “Live beta”, launching the service as a main catalogue, occurred in September, with students being training on its use and providing feedback. We continue to act on feedback, and will go on rolling out enhancements on a monthly basis for the time being.

Wednesday, 20 October 2010

Who are we, anyway?

How remiss of me not to introduce us at the beginning of this project! The main swwhepsrch project team consists of three people all based at Swansea University library:-

Mark Hughes is project head. Mark is Head of Collections at Swansea University and was formerly Virtual Academic Library project manager for SWWHEP, where he was involved in the original choice of VuFind as the project’s discovery platform. He has worked in academic and public libraries for 14 years, mainly in the systems and service management fields, including LMS and RFID implementation. He is leading the work of the case study project.

Paul Johnson is project manager, and is the current SWWHEP Virtual Academic Library project manager, having previously been the project’s systems officer. He has worked in public and academic libraries, and spent much of his career working for library systems suppliers, in software development, data conversion, system implementations and support. He is responsible for the day-to-day management of this project.

Luke O’Sullivan, the project officer, is systems officer and the main developer for the Virtual Academic Library project. His previous roles included five years at Swansea public libraries, working on projects including the procurement, configuration and support of a new LMS and RFID system. He has done most of the enhancement work for VuFind.

Paul