I had mentioned del.icio.us in some meetings last week at work, because I am a strong believer in their “ad hoc taxonomy” approach (which allows end users to think about classification after the data has already been entered, not before where it will raise the bar for data entry). As it turns out, Yahoo! seems to agree: they just bought ’em. Genius move for them, as del.icio.us willl integrate nicely (philosophically as well as technically, one hopes) with their previous purchases Flickr and My Web.

Note that I don’t believe that this means formal taxonomies are useless or pointless or in any way inadvisable: quite the opposite. They are necessary on one end of the spectrum (e.g. the Enterprise Portals of the world) where structured information is a must. However, they are generally too complicated for the average user who just wants to send an email or post a document, which means that a rigorous, structured taxonomy is actually a significant barrier for data classification. Users will prefer to use a collaboration mechanism that doesn’t require a taxonomy, and also unfortunately doesn’t have any public way to perform searches on the data across users: Outlook.

I believe that there is a standard bell curve on this: along one extreme, rigorous taxonomies with strict data classification that requires its adherents to fully grok both the data they’re putting in and the *whole* taxonomy (not just the little bit they are using, otherwise how would they know it’s in the right place?). Along the other extreme, completely unclassified data with no taxonomy, no useful metadata, and no search/indexing capabilities. The problem? Unfortunately, because we don’t currently implement any tools that hit the middle of that bell curve, almost *all* of our data is ending up on this extreme: un-indexed, un-searchable, un-reachable by anyone save the original data creator, in an Inbox, a home folder or a SharePoint site only a handful have access to.

In the middle, there are less rigorous taxonomies that are user-defined in an ad hoc fashion, similar to the way del.icio.us does it. The user defines the tags that are useful and significant to them, selecting not only from their own classifications but from the classifications that the masses have associated with the same or similar data. This “mob-developed” tagging definition (call it “mogging” or “mobtagging” to give it a nice trendy neologism) does two things: (a) it reduces the amount of work required to tag/classify data, which makes it more palatable to the user, and (b) it actually demonstrates to the user the benefits of a taxonomy or tagging system because they are using it directly on their own data. They participate in the taxonomy and the data classification without thinking about it, because (and here’s the important part) the tags are public knowledge.

However, even with del.icio.us I believe there’s something missing: a human eye above the morass, gently nudging the tags in one direction or another. I’m not talking about just fixing typos: it’s about noticing that particular links and particular content and particular tags are associated, so it would behoove the company to tag other, related links with the same tags, suddenly making them available via the search and tagging terms that the users are already using. This is something that is not feasible at the internet level, but is definitely achievable at the enterprise level.

Of course, there’s no one tool to get there immediately, and I don’t really believe that this “human eye” concept is automatable using today’s technology anyway: maybe Google has something in the works (and in fact, one could argue that Google Base is a step in this direction). However, the key to all of this is collaboration, indexing and search, and integrating these things across all the tools that end-users use to publish their information. It’s why you’ll constantly find me ranting and raving about collaboration and publishing information much further than we do today, in ways that the users (not the I.T. people) find easy to manage.

Indexing: X1 Enterprise Edition (indexes and searches across file stores, SharePoint, email)
Collaboration: SharePoint Portal, del.icio.us, Groove, Outlook, blogs, Flickr

Rant over. For now.