IBM’s Taxonomies and Comparing Knowledge Systems (Notes)

Editor’s note: this is a new format for topics in which I want to aggregate and pass along information, but don’t have the time to do a full writeup. I’m calling the format “notes”. These posts will be about a single topic or two and will contain several links to resources that I find interesting along with a bit of contextual commentary. I hope you find them interesting, too. (most Notes won’t be this long)

I’ve been talking a lot lately about the theory behind controlled vocabularies. Most of my talk has been around the idea that they are not inclusive, don’t allow for change, and aren’t as user-centric as folksonomies, the new kid in town.

However, as two commentors politely pointed out, I’m not really giving controlled vocabularies their due. What I’m really doing is pointing out that controlled vocabularies are really hard to implement, which doesn’t exactly mean that they’re bad, but only that most don’t live up to their potential. I’m now convinced that if they were implemented successfully, with a dedicated team, that some of the issues I’ve brought up aren’t as big a detriment as I’ve painted them. (I’m really glad that the folks who comment on bokardo are really bright people)

One of those commentors, Bud Gibson, pointed out an interesting session he attended at IASummit 2005. (My boss Jared suggested that I go next year: he really enjoyed it). Anyway, the session Bud pointed out was given by a group of folks from IBM and was called: Traversing the Corporate Web: IA and Taxonomy at IBM. (here are the notes in .ppt format that didn’t work on my Mac). After combing through the copious notes (presentations are so much better than notes), I came away thinking “Wow, IBM’s intranet is huge!”. Also, I found the team’s problem statement very enlightening:

1. Confusion about how taxonomies work to solve information problems
2. No best practices for using and applying taxonomies
3. Disparate groups developing taxonomies
4. Process-heavy change management around taxonomy
5. No clear engagement path for users & stakeholders to affect change
6. Decentralized taxonomy tools and inconsistent taxonomy formats
7. Strict rules for terminology management that often don’t meet user’s needs

When considering these problems, remember that this taxonomy has over 3,700 categories! Pretty impressive, huh? Also, Bud’s own post about this session included the claim that IBM will be trying out folksonomies to help with these issues, but I didn’t notice that word in the notes…again, presentations are so much better than notes. Folksonomies or not, this team is dealing with issues that are pertinent to all designers.

Continuing on this theme, David Weinberger has written a report for Esther Dyson’s Release 1.0 called Taxonomies to Tags: From Trees to Piles of Leaves, in which he talks about tree-like knowledge structures vs. other options. The report costs a robust $80.00. For those of you without the momentum to pay for that, David has luckily provided the first chapter as a teaser. He’s got a bunch of good insights here, but I’ll only post his interesting conclusions, comparing trees (hierarchies), faceted systems, and tagging systems (folksonomies):

1. Because they are unambiguous, trees work well where information can be sharply delineated and is centrally controlled. Users are accustomed to browsing trees, so little or no end-user training is required. But trees are expensive to build and maintain and require the user to understand the subject area well: How do you find the recipe for bread soup if you don’t know to look in the “Tuscan Cooking” category?
2. Faceted systems work splendidly where an application is being used by such a wide range of users that no one tree going to match everyone’s way thinking. They are also easier maintain than trees because adding new item requires only filling in the information about facets, rather having make decision exactly which category it should go into.
3. Tagging systems are possible only if people are motivated to do more of the work themselves, for individual and/or social reasons. They are necessarily sloppy systems, so if it’s crucial to find each and every object that has to do with, say, apples, tagging won’t work. But for an inexpensive, easy way of using the wisdom of the crowd to make resources visible and sortable, there’s nothing like tags.

So, I hope these prove interesting. Thanks to Bud and David, this has been the inaugural set of Bokardo Notes…