For much of my career I’ve been exploring ways to bring people and information together in shared online spaces. Groupware, social software, the semantic web, the giant global graph, it’s all the same to me in one fundamental way. When we push information into shared spaces, data finds data, and people find people, and all sorts of magic happens.

What do all these acts have in common?

- I publish a blog entry about a technique I just learned.

- I geocode a photo on Flickr.

- I curate a list on del.icio.us.

- I add to a Wikipedia page.

- I arrange a dinner using Windows Live Events.

These all begin as acts of personal information management. In another era I’d have written down the technique I learned in a private journal, put the photo in a shoebox, kept the list in a notebook, scribbled in the margin of my encyclopedia, and recorded the dinner engagement on my kitchen calendar.

When I instead perform these acts of personal information management in shared information spaces, they continue to serve all of their original purposes. But amazing new possibilities arise. Somebody may thank me for blogging the technique I learned, or even better, show me a refinement I hadn’t thought of. My geocoded photos can cluster together with others, revealing unsuspected and delightful connections among people and places. Other people help me grow my list. My contribution to Wikipedia feels virtuous. My dinner arrangements can include unplanned invitees in a context-preserving way.

What do we call personal information management when it moves into shared online spaces? I asked myself that question, and the answer that came back was: social information management. That didn’t seem like a familiar term. But when I searched, I found that it was claimed by my 2005 InfoWorld column, WinFS and social information management.

The questions raised there, about the tug-of-war between formally structured and organically grown information, apply equally to the emerging crop of semantic web applications. A new question now arises about what the organizing structure will be. A knowledge taxonomy? A social graph? A hybrid of the two?

I don’t know the answer, and I don’t think anybody does. But everything that’s happening traces an arc from personal to social information management. From now on we’ll all be learning how far along that arc we wish to travel, weighing the risks and benefits of participating against the risks and benefits of not participating.

Update: Reading Brian Jones’ note about the upcoming XML 2007 conference in Boston reminded me that I once delivered a keynote at that conference. The year was 2003, the title of the talk was The social life of XML, and it was an extended variation on this same theme. For example:

Documents, including the purchase order and the messages related to it, aren’t just passive carriers of information. They’re the warp on which we weave a socially constructed reality. Somehow, we need to find ways to connect that reality to the workflow and process orchestration systems now being invented.

And:

The emerging web services network is radically open — not only because the messages exchanged on that network are XML, but also because the services are connected using pipelines. We can inject intermediaries into those pipelines; the intermediaries can observe and act on the messages. So we can acquire a lot of useful context, and can implement useful policy, by reading and writing what goes by on the wire. Things don’t tend work the same way on the desktop, but maybe they could. Our personal productivity tools are in a position to learn a lot about how we interact with remote services, communicate with other people, and manage our data. And they’re in a position to help us do those things more effectively. But the messages and events flowing on our local machines have nothing in common with the messages and events flowing in the cloud.

At the time I was very excited about the XML intelligence that had recently been injected into the Office applications. And I still am. The use of data-enriched documents in software-plus-services scenarios is growing — slowly, to be sure, but steadily.

If I were to give another talk today, though, I’d want to elaborate on what Adam Bosworth said in his keynote at that same conference. He talked about RSS, about navigating a linked web of data, and about what today I would call (thanks to Rohit Khare) syndication-oriented architecture. Enabling regular folks to produce and consume structured and data-enriched documents is something that will happen gradually, as tools and infrastructure settle into place. But enabling regular folks to produce and consume feeds, and combine them in navigable ways, is something that could happen explosively. So far as I can see, the barriers are more conceptual than technical.