I worked for a number of years in the Content Management Culture group at the BBC. The group was set up by Tom Loosemore’s original Applications team and was spearheaded by Paula Le Dieu. It was also the project that brought Margaret Hanley to the BBC and really kicked off IA at the BBC.
When I joined the project they were half-way to rolling out a CMS to the BBC’s local websites. I spend the first 6 month or so concentrating on the metadata system and then picked up content modelling from Sandra Green. I handled the content modelling for the CMS roll-out to the CBBC site and saw first hand how error prone and labour intensive the modelling process was. The IA created a model in a word doc and then handed it over to the tech team to implement. The implementation in XML was a fairly tedious task for the technical team and working from v. long word documents wasn’t likely to encourage accuracy. So the idea for the COM tool was born.
We just wanted a tool where the IA could input the content model and automatically generate the XML. We carried out a procurement process but ending up settling on adapting an open-source ontology modelling tool, Protege.
This is what the interface ended up looking like. The example shows the model for a page about a book, including required properties, optional properties and containers. A property essentially mapped to field in the user interface. ‘Containers’ were for reusable groups of fields.
The tool actually resulted in slightly more work for the IA as it required us to provide all the information necessary to generate the XML, whereas previously the developers would have filled in some gaps for us/made assumptions. But with greater responsibility came greater power and we could now be confident that what we had modelled would (necessarily) be what the interface ended up with. And now the developers were freed up from creating new XML for every object.
For more information about the implementation of Protege at the BBC see Content modelling at the BBC using RDF and OWL