I attended a workshop held by the Digital Library Federation's Services Framework Working Group, held November 7, 2006. Finally, my kinda initiative. Libraries, technology, architecture, SOA, all combined into one nifty group working together for the common good of mankind. I get the warm fuzzies just thinking about it.
The workshop was essentially a pre-conference for the DLF fall forum 2006, that was held last week in Boston.
My impression was that the working group has done a lot of good work thus far, and I'm very pleased that a group is looking at bringing architecture principles to libraries on a more global scale.
My experience with enterprise architecture and SOA within one library/enterprise -- namely CISTI -- has been quite successful by all accounts. However, that success has been largely won based on a huge communication and education effort, by accepting risks and uncertainties, and by an unending effort to garner buy-in from both management and staff at all levels. I'm concerned that the Services Working Group has it's work cut-out for it in attempting to achieve a more universal acceptance of what is essentially a high-level library architecture effort.
The main criticism I would posit is that if you're doing an effort such as this, to capture the common business processes of libraries, then please don't make up a process or methodology. The skills, training and methodologies exist to do enterprise architecture (or, perhaps in this case - meta-enterprise architecture).
Also data flow models are not passé. Don't make things complex that can be simple. Most times I find that if you can identify the business process, data, data flow and actors... you've solved your problem. :)
A framework should be developed at a particularly high level, encompassing only the common and agreed upon elements of library processes. Whilst you may need to dig deep to collect and confirm processes, the framework itself, I suggest, should remain fairly high -- providing individual enterprises the ability to compare, contrast and build upon that framework in their own context. That said, libraries have been around for a very long time, I'm certain that libraries have many business processes that they commonly share.
What am I saying? I'm saying there are at least 2 levels of architecture here. The high level meta-architecture (framework) thats generally agreed upon amongst libraries, and then there's a true enterprise-level architecture that's needed within an institution to meet specific needs. The enterprise-level architecture should, ideally, use the framework to guide their architecture development and implementations... but a framework can never fully accommodate the specific business needs, planning and implementation required within an organization. That said, it was very clear from the paper published in DLib Magazine and during the workshop that the Services Working Group in no way was attempting to delve too deep here. What I'm saying is that libraries should get some architects on board and within their planning and technology groups, and perhaps the Services Working Group could attract some architects to assist them in their valiant efforts.
Another thought: Who the client/user is for this architecture. The client, ultimately, is the patron -- not the librarian, I think. Is this true? I always find it easier to attack these questions from the end-user perspective first. Argue with me.
Overall, I believe that it is a timely, well-needed and important effort. I'm hoping for its success.
For more information please consult the Digital Library Federations Services Working Group page, and the Education Commons DLF Services Framework site.
Check out some of my photos around Boston that I took during the workshop. It's a beautiful city to inspire thought.