OSS Watch's  second national conference focused on an often articulated anxiety concerning how an institution will answer the question of support when considering the deployment of open source software. OSS Watch is a pilot advisory service set up by the Joint Information Systems Committee (JISC) to provide UK Higher and Further Education with neutral and authoritative guidance about free and open source software. Whereas OSS Watch's inaugural conference in December 2003  presented an overview of the entire field, this event concentrated on what is sometimes thought to be the single most significant barrier to institutional take-up of open source software.
Sebastian Rahtz, Manager, OSS Watch
The first session of the day, chaired by Paul Browning, University of Bristol, started with a presentation from Sebastian Rahtz. Sebastian set the scene by exploring some of the data gathered in OSS Watch's scoping study  on open source development and deployment in UK Higher and Further Education. That study revealed that 37% of respondents (41% in the case of further education colleges) cited support as a major concern. Such a reservation might be a very practical barrier to open source deployment. Moreover, this could be a significant differentiator between open source and propriety software. It could be the deciding factor when an institution is engaged in a procurement or deployment decision-making process. But is this anxiety in fact well-founded? The four models to be explored throughout the day do not exhaust the range. They suggest that there is a spectrum of support options available. But there is no silver bullet. The surprise here, if there is one, is that the exact same thing can be said of support for proprietary software. Increasingly open source and proprietary options share the same ground, e.g. IBM or Novell support both, there are training companies who support both, there are books about both, help on the Web about both, and so on.. The difference, where there is one, returns to the rough ground - fitness for purpose. And that, Sebastian suggested, is probably the best basis for choice that anyone could recommend. If open source is up to the job, or does a better job, that should be the reason it is chosen. Support, perhaps, is, or is becoming, merely a co-factor.
Adam Marshall, University of Oxford and Joel Greenberg, The Open University
Adam Marshall  gave a personal account of his involvement in the development of the Bodington  Virtual Learning Environment (VLE), as deployed at the University of Oxford. Bodington was developed in-house at the University of Leeds and was later 'open-sourced'. That decision process itself might be worthy an independent case study. Adam, however, concentrated upon Oxford's direct involvement in the project of taking Bodington development forward. Bodington is a relatively small project in terms of participants (3 or 4 big sites); developers in Oxford therefore have close contact with its originating developer, Jon Maber. Indeed, Adam noted that such contact was invaluable. Amongst the difficulties in this deployment was the effort to convince funding bodies that money spent on development equated to money that might otherwise have been spent on a yearly licence for a proprietary VLE. However, there have also been unexpected benefits. With sufficiently experienced Java programmers now working on Bodington's development, Oxford has also been able to bid for other projects in the e-learning sphere and elsewhere; which taps an otherwise unreachable funding stream for ongoing development.
Joel Greenberg approached the DIY theme from the institutional perspective. His presentation, Deploying Open Source Solutions in an eProduction System , demonstrated how, from the institutional perspective, there is little difference between open source and proprietary software. From this more global view the issues tend to be those of interoperability and workflow. Joel explained how The Open University had arrived at its eProduction environment. This environment uses a mixture of proprietary and open source solutions moving from a localised implementation of Word 2003 for structured authoring, through an in-house tracker system and content management server, to XML conversion and later preview presentation as HTML via Apache Cocoon. The choice at each step between open source and proprietary software was based strictly on fitness of purpose. Where in-house solutions worked best, they were chosen; likewise when proprietary solutions worked best. Pragmatism (and good practice), not ideology, were the driving factors.
Ian Dolphin, University of Hull and Paul Cooper, OpenAdvantage
The second session of the day was chaired by Andrew Savory, Luminas. Already in the DIY session there had been considerable discussion about linking with other institutions to share solutions and development work. Ian Dolphin's presentation, JA-SIG & uPortal: The Hull Experience , confirmed that in some cases such as institutional portal deployment, there is much to be gained through joining a consortium. The consortium in this case was the uPortal project of the Java Architecture Special Interest Group (JA-SIG). Players within this group include Yale University, Columbia University, University of British Columbia and other large North American institutions. Ian noted that, among other reasons, it is certainly no bad thing for The University of Hull to be mentioned in such company due to its involvement in the consortium. This raised an interesting point. Deployment decisions may have ramifications at many removes from their direct application. Sometimes decision-makers need to step a long way back in order to see the big picture.
Paul Cooper of OpenAdvantage knew the practical benefit of building consortia. OpenAdvantage  is an independently funded vendor neutral organisation bringing together small and medium size businesses in the West Midlands with public sector institutions around the use and support of open source software. Consortia can be large international ventures or they can be very local. The importance is to leverage the different skills within the group for the advantage of all.
John Merrells, Parthenon Computing and Michael Sekler, OS Consult
The third session of the day was chaired by Mike Banahan, GBdirect. Both presentations for this session involved open source consulting companies explaining what they are and what they do. To some this may have sounded like a sales pitch. That would have been no bad thing if it were true. It is time to realise that small businesses are springing up to support open source deployments. John Merrells  drew a distinction between direct and indirect support. For some, support is about solving the 80% of deployment issues that prior experience with the software makes plain. For others, support involves the need for the development of special features in the software. Open source deployments, of course, make that second case far more practical. Small consultancy firms often seek to cover both consultancy needs. Michael Sekler's  firm is primarily focused on that second group. They work closely with Wynona, the developers of Apache Lenya, a content management system based on the Apache Cocoon framework.
Simon Lidgett, Novell and John Heath, Sun Microsystems
The final session of the day moved open source support onto an entirely new level. Chaired by Tony Brett, University of Oxford, this session focused on the support for open source software provided by substantial corporations. Simon Lidgett  of Novell began with an overview of the Novell solution for educational institutions. Novell's solution offers a mixture of open source and proprietary software that builds a complete stack suitable for solving most needs. Interestingly, the Linux operating system is a substantial component within the stack. Novell do not treat open source as peripheral to their support packages. Open source software is as fully supported as Novells' own proprietary software. Of course this kind of support comes at a price. But that too is consistent with the spectrum view of open source support.
John Heath of Sun took this opportunity to outline a new initiative within Sun for the education community: the Java Education and Learning Community (JELC) . Sun is fostering a community to share Java-related open source educational tools, open learning standards implementations and open course learning materials.
Support Models for Open Source Deployment was a lively day full of debate and discussion. Sebastian probably got it right in his opening remarks: there is no silver bullet. Support for open source is a broad church and there are many different models available. What is most clear, however, is that the concern that there is no support available for open source is certainly unfounded. Of course this does not mean that open source is necessarily on the same footing with proprietary software in every sphere of deployment. Decisions, as ever, need to be based on the reality of what is available at the time. But that begins to sound rather like simple, practical, advice that would be true whether you were thinking of deploying an open source or a proprietary solution. Perhaps the fact that this comes as no surprise is itself the most surprising observation that could be gathered from this conference.
Sebastian Rahtz is Information Manager at Oxford University Computing Services, and is also manager of OSS Watch. He is also on the Board of Directors and Technical Council of the Text Encoding Initiative Consortium, and author of various open source packages in the XML and TeX worlds.
Randy Metcalfe is communications manager of OSS Watch, a role he previously filled for the Humbul Humanities Hub.