Knowledge micro-management

I’ve been thinking about Paul Ford‘s notes on knowledge management. I don’t think I agree with Paul’s assertions that “individuals solve an enormous amount of knowledge management problems on their own”, and that “the best KM would observe how people actually organize data on their desktops in an organization and allow them the same degree of flexibility with Organizational Memory data”. This isn’t knowledge management, it’s knowledge micro-management. The whole field of KM is extremely imperfectly understood, and the first guy who comes up with a decent solution, assuming that it’s possible for one to exist, will be a millionaire. But I’m not at all convinced that the best way to handle KM is to enlarge individuals’ ad-hoc KM techniques into an organisation-wide scheme. That’s knowledge micro-management, as I say. Now, it’s possible that what Paul meant was that people store their data how they like but the KM system understands everyone’s storage rationale and categorises the data correctly underneath. If that’s the case, then I think it’s a good idea, except that I can’t see how it will work. I admit that I’m not convinced of the current crop of KM tools either, since they seem to work on a principle of devising one schema for data and making everyone stick to it, with the inevitable result that everyone thinks it’s wrong and no-one uses it properly. This is still clearly better than the no-KM method, which involves everyone having to understand how everyone else’s personal data management rationale works and just does not happen at all, with the result that there’s loads of knowledge out there that should be being used effectively and isn’t. Paul’s solution might be a silver bullet — store your data how you want and the system, learns about and understands your method, extracts the data from your storage and categorises it itself. Ideally, it could then reflect that data back to everyone else, using each person’s individual data storage preferences; so you perceive the system as working solely for you, when in fact it works for everyone. Great idea. But it’s a pattern recognition problem, recognising what a bit of data is and where it belongs, and pattern matching is something that computers are very, very bad at. The best approach to that sort of thing that anyone’s come up with so far is in things like the Insidious Big Brother Database in Emacs’ mail client, where every bit of data is tied up with similar bits — an email address can lead you to other mails from that person, or other people from that mail domain; a subject can lead you to other mails in that mail thread, or other messages sharing keywords with this one. Essentially, you present a user with all the information and then let them sort through it to find what they need, because people are very good at pattern matching. This is the approach that KM seems to be heading for; you leach all the information out of someone’s head and then drop it into a massive data warehouse, and then categorise the data according to a given data schema. It may not be the right approach, but I don’t think that Paul’s approach, while initially more tempting, is possible given current technology — and a poor implementation of it would be worse than no implementation, because a bad KM system gets in everyone’s way and makes you jump through hoops to accommodate it. The ideal KM system would be one that users didn’t even know about; it quietly grabs and categorises data and makes it available, all without you knowing that it’s there. Paul’s method is a step toward that ideal, but it’s not a doable step. ——-

I'm currently available for hire, to help you plan, architect, and build new systems, and for technical writing and articles. You can take a look at some projects I've worked on and some of my writing. If you'd like to talk about your upcoming project, do get in touch.

More in the discussion (powered by webmentions)

  • (no mentions, yet.)