A DRY CRIG Event for the IE Demonstrator
In June this year UKOLN hosted an ‘unconference’[1] which was given the title ‘DRY/CRIG’. Jointly funded through the IE Demonstrator Project [2] and the Common Repositories Interfaces Group (CRIG) [3], this event was intended to allow technical representatives of (mainly) JISC-funded ‘Shared-Infrastructure-Services [4] to meet software developers from UK Higher Education institutions (HEIs). The ‘DRY’ part of the name is an acronym standing for ‘Don’t Repeat Yourself’, a general principle in software engineering, which was deemed appropriate for an event mostly concerned with reusable shared services. It also incorporated a slight pun, as the CRIG DRY events have tended to be held in bars!
For the first part of the day, the service representatives had been briefed to deliver a five-minute ‘elevator pitch’ for their service. Because their audience had a high degree of technical knowledge and experience, the speakers were encouraged to deliver this pitch in more technical terms than they might have normally. The idea was for the speakers to try to persuade the HEI developers to engage with their service at a technical level. The second half of the day was devoted to the sort of prototyping activity which has come to characterise CRIG events.
The event was jointly organised by David Flanders (representing CRIG) and myself. We had previously agreed to organise a CRIG event in Bath; as the academic year had finished, I found that the Parade Bar [5] on the University of Bath’s campus was available to use as a venue for the 20-30 people we invited. This turned out to be an excellent venue with plenty of space, free WIFI provided by the University, food and drink available all day at the convenience of the delegates, comfortable sofas and even a pool table!
Following introductions by David and myself we jumped straight into the ‘elevator pitches’ session.
STAR Project
Ceri Binding, University of Glamorgan
Ceri Binding of the STAR Project [6] gave an excellent overview of the outputs to date of this effort to develop services around semantic technologies, initially within the subject domain of archaeology. Ceri explained how the STAR Project has already done some of the ‘heavy lifting’ to prepare vocabularies for use in semantic applications, such as presenting Library of Congress Subject Headings [7] as SKOS [8]. With the provision of some publicly available Web services, this work could then be exploited by institutional developers. The particular use case which Ceri ‘pitched’ to the audience was that of a service providing functionality to support the expansion of searches being conducted by someone using an institutional system such as a repository.
GeoCrossWalk
James Reid, EDINA, Edinburgh University
GeoCrossWalk [9] offers a ‘spatially aware’ database, together with a Web service-based API and a ‘geo-parser’. Together, these elements make up a package which allows information to be parsed for geo-spatial terms. Such information can then be given a new, coherent geographical context. With the widespread use of Google Earth [10], it is apparent that such services are attractive to users. The potential application to repositories and other institutional systems is interesting, and James’s pitch generated a number of questions from the assembled developers. The chief concern for many is the restriction placed on the use of this service because of issues around IPR: the source data used for the main service is licensed from the Ordnance Survey (OS) [11] and so cannot be made openly available. EDINA are negotiating a new licence with the OS, due to come into effect in the Summer of 2009, which will give access to the data to organisations in the UK academic domain. The audience were, however, not convinced that even this would be open enough.
While the current dataset underpinning the GeoCrossWalk service is contemporary in nature, an understanding has been reached with a supplier to provide data similar in type but historic in scope.
Names Project
Daniel Needham, MIMAS, Manchester University
Daniel introduced the Names Project which is developing a names authority service for academic authors and organisations. Using the approach of mining a wide set of data sources and then de-duplicating and disambiguating the results, the intention is to provide a centralised, authoritative database of uniquely identified individuals and organisations. Eventually, each unique name will be given a persistent identifier, possibly using the Handle system [12]. Daniel presented possible applications for this in an institutional context, such as in the workflow around the deposit of a scholarly work into an institutional repository, where the ‘author’ field in the new item’s metadata record could be validated against an authoritative source.
Although Daniel could not demonstrate any running software, the intention is to have a prototype ready for testing in September 2008. When Daniel revealed that the service was being developed in Java with a SOAP [13] interface, the audience of developers was unanimous in its plea for a simpler, URL-based or ‘RESTful’ [14] interface.
Among the issues which were discussed were privacy, security and long-term preservation. Features which may need to be developed and offered include ‘self-management’ – where the subjects of the name authority entries are given the means to manage their own record, and some sort of ‘dispute resolution’ mechanism.
HILT Project
Anu Joseph, Centre for Digital Library Research, University of Strathclyde
Anu introduced the HILT system which is intended to improve interoperability in cross-searching and browsing of structured information. Essentially, it is used to map between controlled vocabularies. In the context of an institutional repository, it could provide services to the deposit workflow – allowing the selection of alternative terms to be entered into metadata records. At the ‘other end’ of the repository, HILT could assist in providing a richer search interface by offering query expansion services.
Like the STAR Project, HILT is actively developing SKOS-based interfaces. Anu’s pitch was followed by questions from the developers, mainly about how useful SKOS can be in satisfying machine-machine requests for information.
Deposit Plait Project
Stuart Lewis, Aberystwyth University
Stuart introduced Deposit Plait [15] – a recent JISC-funded interoperability demonstrator project. This is concerned with investigating, and possibly (but not necessarily) prototyping a ‘deposit engine’, designed to reduce the amount of metadata which is required to be input, manually, into repositories. Initially, the project will attempt to establish a ‘base set’ of metadata, before exploring solutions for aiding the creation of this metadata through the exploitation of remote services as well as the processing of the resource being described by the metadata.
Foresight Project
Rob Sanderson and Richard Jones (HP Labs)
Rob Sanderson briefly described their work with the OAI-ORE [16] specification as part of their Foresight Project [17]. Rob gave a useful ‘high-level’ description of ORE, before joining Richard in describing the software libraries they have developed in both the Java and Python programming languages. They invited the developers before them to download these libraries with a view to examining and testing them. This invitation met with a positive response.
CRIG Prototyping Session
Following the succession of five-minute pitches we held a general discussion before breaking for lunch. With a good range of hot bar food available lunch was, by all accounts, a success – proving more popular than the traditional sandwich buffet.
After lunch we moved into the CRIG session, where smaller groups were formed around particular ideas and a period of prototyping ensued. I joined one group discussing ideas around the use of GeoCrossWalk to provide a service supporting the ‘faceted’ browsing of information based on geo-spatial classification. We explored the fascinating notion of providing such services in a ‘just-in-time’ fashion, rather than applying such metadata to huge quantities of information in anticipation of a use for it.
As evening approached, the majority of delegates left, leaving a smaller group of developers to enjoy a drink and a more free-flowing conversation. Later that evening, when I overheard one or two terrible jokes involving HTTP response codes, I concluded that the useful part of the day had ended and that it was time for me to go home!
Conclusion
In conclusion, and judging by the feedback I have received, the event was a success. The use of a bar worked very well, except where it got a little noisy towards lunch-time, while the menu of hot food was widely appreciated. The objective of getting technical service providers together with institutional developers was met, and the discussions which flowed from this were characterised by mutual learning. In many respects, while having a clear purpose, this workshop was also an experiment in event organisation. The use of a bar meant that delegates could order refreshments at their own convenience, rather than waiting for the tea and coffee to be delivered as is usually the case. Moreover, the cost of the event was modest: allowing delegates to select from a varied menu of hot bar food was marginally more expensive than providing them with a basic lunch of sandwiches, but the deal we struck with the bar management meant that this was the only direct cost. Although I had reserved a more traditional teaching room in the University just in case the bar became too noisy, the gamble of using what was, essentially, a public space, paid off. Having tried this approach, I would be happy to run an event in this way again.
Several people have commented publicly (and positively) about the event [18][19][20].
As one of the organisers, I would like to extend my gratitude to all the people who turned up, participated, and helped to make the event a stimulating one. We intend to organise at least a further two similar events as part of the IE Demonstrator Project – if you think that you might be interested in participating, keep an eye on the IE Demonstrator blog [2] where an announcement will be made. I would especially like to thank David Flanders and Mark Dewey (UKOLN) who helped organise and run the event.
Acknowledgements:
I would like to express my thanks to Andrew McGregor of the JISC Information Environment Team whose notes have proved an invaluable asset in the writing of this article.
References
- Definition of an unconference http://en.wikipedia.org/wiki/Unconference
- IE Demonstrator Project http://www.iedemonstrator.org/
- CRIG http://www.ukoln.ac.uk/repositories/digirep/index/CRIG
- JISC SIS http://www.jisc.ac.uk/whatwedo/programmes/programme_rep_pres/shared_services.aspx
- Parade Bar, University of Bath http://www.bath.ac.uk/hospitality/parade/
- STAR Project http://hypermedia.research.glam.ac.uk/kos/STAR/
- LCSH http://hypermedia.research.glam.ac.uk/kos/STAR/
- SKOS http://www.w3.org/2004/02/skos/
- GeoCrossWalk http://www.geoxwalk.ac.uk/
- Google Earth http://earth.google.com/
- Ordnance Survey http://www.ordnancesurvey.co.uk/oswebsite/
- Handle system http://www.handle.net/
- SOAP http://en.wikipedia.org/wiki/SOAP
- REST http://en.wikipedia.org/wiki/REST
- Deposit Plait http://www.jisc.ac.uk/whatwedo/programmes/programme_rep_pres/interoperabilitydemos/depositplait.aspx
- OAI-ORE http://www.openarchives.org/ore/
- Foresight Project http://foresite.cheshire3.org/
- OUseful blog http://blogs.open.ac.uk/Maths/ajh59/014735.html
- IE Demonstrator blog http://blog.iedemonstrator.org/2008/06/07/crig-dry-workshop/
- Ian Ibbotson’s workblog http://k-int.blogspot.com/2008/06/jisc-crig-iedemonstrator-barcamp.html