There’s a discussion on a Data Management discussion group (dm-discuss) at the moment, which started on the topic of the different levels of data model (see the pyramid), then moved on to the ‘dream’ of metadata repositories. I said that I know of a company where the business use a metadata repository, and their IT department has nothing to do with it. Am I misusing the word ‘metadata’ here? Is metadata just ‘data about data’; does IT have a monopoly on the term?
I first came across a model of metadata in the early ’80s. We didn’t call it metadata then, but that’s what it was. It was the metamodel for the ICL Data Dictionary System (DDS), and it had four quadrants. One axis was Business vs. IT, the other was Data vs. Process. Some element types (like Data Item, I think) sat on one of the boundaries. As a result, I’ve never assumed that metadata was just IT stuff. Pete Rivett at Adaptive is possibly the author of that metamodel, as he was working on the ICL DDS at the time. In the Wikipedia entry on the ICL VME operating system, there is this nice statement of the role of the DDS:
“an early and very successful attempt to build a comprehensive repository supporting all the other tools, with full support for the development lifecycle. As well as database schemas and file and record descriptions, the dictionary keeps track of objects such as reports and queries, screen designs, and 4GL code; it also supports a variety of models at the requirements capture level, such as entity-relationship models and process models.”
I have a simple diagram that illustrates the breadth of metadata at a high level – see http://metadatajunkie.wordpress.com/metadata-archipelago/. I’ve worked with Adaptive and Rochade in the past, and they both have clients whose main use of the repository is to manage business artefacts. A Dutch company called Resilient sell a Project Portfolio Management product, which utilises the Adaptive Repository; the unnamed company I referred to above may be using their metadata repository to manage their strategies, organisation structure, processes, who knows what.
In the end, it’s all data about how the business operates; it doesn’t matter who looks after it, it’s all metadata to me. Ideally, of course, they ought to recognise that the organisation as a whole needs to integrate their metadata, stop creating more islands in the metadata archipelago, start building bridges, and merge the islands into a ‘continent of metadata’. OK, perhaps I’ve taken the geography analogy a little too far, but I hope you forgive me.