Beyond Lifestreams: the inevitable demise of the Desktop Metaphor

Search has become the holy grail of the desktop. The predominant search companies are moving into desktop search as a way to extend their reach and operating systems makers are moving aggressively in this direction as well. Integrating search into our desktop environments is not a new development - Microsoft has been pursuing WinFS in fits and starts over more than the last decade - but this time something is different: everyone thinks it's important.

Why is there a growing interest in search on the desktop? Many attribute this interest to the "googlization of the desktop." The thinking goes: If my desktop is becoming as complex and hard to navigate as the Web, then why not apply the principles that work on the Web to my desktop? It's not a half-bad idea and its early success clearly demonstrates a need for new technologies (and possibly new metaphors) that can help us manage our electronic lives.

But is "desktop search" what we're really after? Will it allow us to finally manage the deluge of information pouring into our desktops? We didn't think so in the early '90s and we don't think so now. Our claim: To fully deal with the problem we need to break from the desktop metaphor and move to a model that removes the overhead and design constraints introduced by that metaphor. Search on the desktop is actually a step in the right direction, and by moving in that direction we've gained something quite valuable: an alternative model for how users might approach taking back control of their electronic lives. In fact, over the last decade, Web search has primed users for a different style of managing their electronic lives. However, as we'll see, we're not there yet; search is a necessary component of such a model; but it won't be sufficient by itself.

In 1994 David Gelernter and I undertook what some considered a radical approach to fixing our electronic lives by creating an alternative to the desktop metaphor. Our approach, Lifestreams [Free][Gel], is a software architecture for managing personal electronic information. The approach was radical in the sense that Lifestreams threw out filenames, folders and static filing and in their place provided a simple data structure: a time-ordered stream of documents combined with a small number of powerful operators for locating, organizing, summarizing and monitoring information. Our prototype implementation at Yale University realized many of the system's defining features and allowed us to experiment with the model's key ideas. Later commercial releases implemented a narrower set of Lifestreams' features, yet met real world usage.

Ten years later one only has to look as far as Apple's "iApps" (not to mention quite a few other efforts) to see many of Lifestreams features in action. While we make no claim that our work directly, or even indirectly, influenced Apple or a number of other software vendors, we have always claimed that the defining features of Lifestreams were natural and would end up in information management systems down the road. That now appears to be happening and as a result there is some renewed interest in our early work. So, in this Chapter we'll take a look back at the early work we did in alternative desktop metaphors.