[ptsefton.com] | [CV & Bio]

The delicious intranet

2004-08-23

The week before last, I discovered the communal bookmarkathon that is del.icio.us via Jon Udell. Since then my team and I have been looking at the possibilities. One possibility: improve our ageing intranet.

The del.icio.us approach is to let everyone classify everything they put on the site (just URL's) in their own terms, in a simple flat free-for-all of keywords. A bit like jelly wrestling (Jell-o wrestling, in some parts of the world), only in keyword-soup rather than gelatine. A system like this would allow our users to whack-up stuff and label it however they want. But the official 'project view' could be maintained by our (fully qualified, dynamic, talented) project manager. And HR people could tag stuff that is relevant to new staff, naughty staff, staff-member-of-the-month type staff, and so on without anybody having to maintain taxonomies.

Is this the solution to the perennial problem with the intranet where the classification system my team set up makes no sense to some users? "Is this a Discussion Paper or a Form or a Procedure?" they wonder. And they want to make their own classifications. Imagine that. "Can I make my own folders?" they ask.

The NextEd Intranet is either a triumph of content management or a @#$^*&#@%, depending on who you talk to, and on what day. It is metadata driven; when you add stuff to it, it wants to know which group owns the content, what type of content it is, a title, and so on. It then decides, in its almost infinite wisdom where to keep your content, to let others browse and search their way to it. But any coercive metadata gathering results in some pollution of the data. For example, it wants an abstract for everything. It's amazing, really how many people can supply an abstract that has less information in it than the document title.

We have been talking for a while about an Intranet that does not coerce people into providing information that may be flawed, but lets them get their stuff in there, and lets the classifiers who are good at abstracting content and organising things for their domain pull it in to shape. Del.icio.us provides a good model.