When trying to integrate diverse vocabularies and repositories, the way to go is mapping – metadata crosswalks as they are known in the US. I’ve been looking for software that can handle mappings between taxonomies, of which there are a range on the market, but what is really exciting is the development of automated mapping tools to take much of the “heavy lifting” out of the work (for example Synaptica’s AutoMatch).
It seems to me that there is a convergence between semi-automated mapping (we’ll be needing human editorial oversight for some time) and the semantic web project. A combination of auto-mapping and RDF/OWL/SKOS should enable us to cross-navigate repositories using our own terminologies. This is the realisation of the “many perspectives, one repository” approach that should get round many problems of the subjective/objective divide. If you can’t agree on which viewpoint to adopt, why not have them all and save the arguments for the nuances of the mapping process. Within organisations this has immediate benefits, in removing a lot of politicking that surrounds information and knowledge management. However, there is also huge cultural potential when it comes to opening up public repositories and making them interoperate. The Europeana project is a good example.
A wise taxonomist once said to me “taxonomies are technology agnostic” and I’ve been thinking about why systems are not taxonomy agnostic. If you underpin a taxonomy with a thesaurus, can you use that to map one taxonomy to another, without altering either taxonomy? You can keep both taxonomies as metadata attached to your asset and expose one or the other depending on user choice. It’s just an interface issue. The mapping would enable cross navigation, so you could wander down one taxonomy, skip to another, then pop back to the first one if you wanted.
You could attach folksonomies too if you wanted to, and just store those as extra metadata.
I can see that there might be terminology issues that need resolving (no small task), or perhaps software or storage issues, but I can’t see why the system itself couldn’t work in theory.
I’ve spent a lot of time thinking about mediating stakeholder needs to get the best taxonomy, and that is still a valid approach when you need management and control, but I don’t see any reason not to attach other taxonomies to your core taxonomy. Those satellite taxonomies can then serve minority interests or specialised needs. As long as you collect metadata about your taxonomies and make it clear to your user the provenance of the taxonomy or folksonomy they are viewing, you can offer a range of viewpoints.
Perhaps I am missing something obvious, but it seems there is still debate about getting the best taxonomy, or choosing to implement one instead of another. That debate seems to be based on the presumption that you can only have one taxonomy at a time, but why not have lots?