I’m in the middle of a very stimulating and uplifting exercise together with some colleagues an two brilliant chaps from IBM’s Hursley Lab. We started the week with a 2 day crash course of the product. Yesterday was spent modeling the objects we want to put into WSRR to implement the Baseline Service Documentation model. In preparation for this we did some homework around updating our existing model to better reflect the experience we’ve had with trying to build truly reusable services, something that the old model only partly supported. Also, we wanted to align us more clearly to WSDL’s, something which is necessary if we want to play with WSRR, since all it understands out of the box is WSDL.
Today we finally started putting the pieces together. We learned about Protégé, a free tool to to edit ontologies. Within a couple of hours, we had a first draft of an OWL-file (defining our model, with objects and relationships). Once imported into WSRR, magical things started to happen. Stuff we’ve been dreaming of, but have not been able to realize, include:
- Automatic dependency analysis, meaning that we can instantly ask things like “If we bring down this machine, what services and service consumers are affected?”
- Automatic name generation, meaning that we can enforce our naming conventions much better than before
- Automatic GUI generation based on the model, complete with links to add related object, so, for example, while creating a Service, we automatically get a link to add Operations
We’ve also seen the relative ease by which you can create java plug-ins to enhance the out-of-the-box functionality of WSRR.
It does look really promising so far! Check back soon, I’ll wrap this up with some more thoughts.