Monday, June 19, 2006
On the Relative Presence or Absence of Shelves
This is a great article from Clay Shirky about how Ontology is overrated - and how the human obsession with cataloging things and building file plans is slowly being questioned as the best way to organize information. My favorite quote is the one about building classification hierarchies requiring two skills that are hard to find: "Mind Reading and Fortune Telling"...
If you're at all interested in information management, or the future of the internet, or knowledge systems, I'd urge you to go and read it right now:
My take is that while the shelves have indeed vanished, and some librarians are wringing their cardigans in denial, that from an enterprise content management position, the de facto tag driven classification schemes aren't quite enough to be able to guarantee the ability to obtain the complete business context of whatever information you've found.
If I need to find a crucial accounting document, I need to know that all the related documents in my system are related and available for review also. For example the emails that led up to the decision to report all the crates of champagne on the expense report as 'Correction Fluid"...
Using an evolutionary ontology like del.icio.us only creates a trend towards the relevant business context being made available - and when you're talking about an entire enterprise, even having a high probability that a critical related record will be related is not good enough. You need to know for sure that all the related records are there.
Until auto-classification, or auto-tagging is a completely reliable process, I can't see a major shift occurring from planned taxonomies and containers deriving from them - at least not for vital records management, anyway.
For the random federated collection of stuff that we have come to depend on every day, well, that's a different story - you can see my loosely tagged and wildly silly contribution to the taxonomy of the entire internet at http://del.icio.us/goodgord