Last night I stopped by the Symposium for Social Architecture, sponsored by Corante and the Berkman Center at Harvard Law School (I hosted one of the panelists, and another one crashed on my couch following a red-eye flight from San Francisco). The “social architecture” up for discussion is really about social software, which has proved to be a very useful term for framing contemporary Internet technology. I was curious about how it should apply for businesses and other community of practice, and whether it is all-encompassing. But first, I wanted to make sure I had a complete understanding of what the term has meant over time.

Thankfully, a year ago, Silicon Valley software entrepreneur Christopher Allen did just that: he posted an extensively-researched, 4400-word article tracing the evolution of social software through various incarnations of conception. According to Allen, the ideas had been “traveling through terms such as Augmentation, Groupware, and CSCW in the 1960s, 70s, 80s, and 90s.”

What’s in popular use today is the definition suggested by Clay Shirky over 2002-2003. At a keynote address for the O’Reilly Emerging Technology conference in April 2003, he offered this quick definition: “software that supports group interaction” (followed by 9,000 words) Shirky explained to Allen that he applied the term to distinguish the emerging technologies of weblogs and other communities from the familiar, and stale, corporate groupware tools.

This type of social software is about social relationships— seeking them, inviting them, enhancing them, and as Shirky wryly refined it earlier this year, consummating them. The name works because it encourages us to think about the social values of a technology. A typical social software might manifest certain social features, like ranking top contributors. This is typically missing from Lotus Notes or any derivative corporate groupware. Or it might give prominent display to user comments. Given that comments were missing from her local newspaper’s blog, reader Tish Grier declared it anti-social. When it has comments, they should be set up in a way to promote social relationships; otherwise, it would be like my experience with the Howard Dean forums two years ago, which I called anti-social. People who share positive social experiences will return.

Nonetheless, Allen traced a different vision of what social software could have been. The idea of a personal electronic library was in Vannevar Bush’s conception of a “memex” in 1945; it was rejuvenated by Internet/pioneers J. C. R. Licklider, Doug Englebart in the 1960’s. It was labelled as “social software” for the first time by scientific visionary K. Eric Drexler, founder of the Foresight Institute. Though the term in this use never quite caught on, Drexler’s ideas were presented at a time when personal computers had started to become mainstays in society. Drexler submitted a paper to the Hypertext ’87 conference entitied “Hypertext Publishing and the Evolution of Knowledge.” It was not relationships that Drexler was interested in finding and enhancing but knowledge. Drexler was fully familiar with the academic publication process, which he summarized as “write-up, then to submission, review, rewriting, resubmission, publication, and distribution” followed by “reading by a critic, an idea for a refutation, write-up.” While this had served the literate world for thousands of years, it was a disaster from an information engineering perspective. The process was too slow; and it would allow mistakes to persist uncorrected in libraries everywhere. Thus Drexler wanted to redesign the whole process of disseminating information. These are functions he saw of social software:

voting and evaluation schemes that provide criteria for later filtering. … for collaborative development of modeling games and simulations … group commitment and action: individuals could take unpublicized positions of the form I will publicly commit to X if Y other people do so at the same time. Once Y people take a compatible position, everyone’s commitment (to making a statement, forming a group, making a contribution, etc.) could be automatically published. The possibilities for hypertext-based social software seem broad.

Drexler’s notion of broad possibilities were a utopian vision of what advanced knowledge would do: generate wealth, lead to better economic policy; improve health care; aid survival (at the time, preventing nuclear war). Not only broad, but grand.

(There were other curious assumptions that Drexler made through the essay: that there would be people who would only write hypertext for a living, and thus the system would require supporting royalty payments; that “hypertext publishing might reach only a tiny minority directly”; that readers could “get lost”; that search engines would not develop; that lacking any advances in telecommunicatons improvements, CD-ROM’s would be the medium of choice; that new or unrecognzed authors will not be able to “pass a typical filter,” though they could get around this by emailing their friends. This last point is not so far-out, once one considers that the “filter” is not really a formal one, but a natural one which arises due to the limitations of attention; I covered this extensively in the New Gatekeepers series).

A decade later, with the World Wide Web finally breathing life into online hypertext, Foresight sponsored a project to meet the original vision of knowledge constructon. The project, CritLink, was developed by Ka-Ping Yee, then an undergraduate at the University of Waterloo and today a PhD student at Berkeley. CritLink was a tool to allow a web user to annotate web documents as they surfed and potentially share them with other researchers using the same tool, and visiting the same sites. Wendy Seltzer, at that time a student at Harvard Law School, started adapting the Crit tools for use by the Berkman Center’s H2O Project for community-based educational software. I was one of those community members who looked to help when the Berkman Center opened up their h2o-discuss mailing list – a gesture of openness on the part of the Center which I ultimately got around to thanking Mr. Berkman himself this past January. On the list, Wendy and I discussed some of the possibilties of the software, and the challenges ahead. I eventually documented eleven such annotation tools, which was a small contribution to the project.

Nine of those tools have been largely abandoned. The biggest commercial venture, ThirdVoice, had the temerity to have its software change the content of a web page on a user’s browser. In 1999, the concept of a “remix culture” was not yet the new religion of the web, and a Say No to ThirdVoice protest group was born. (It didn’t help that this particular type of remix attracted mostly spam.) The simplest system I evaluated was one that didn’t trouble to annotate specific parts of the document; rather it nchored comments to the whole page, which itself was usually not much more than a hyperlink to a news story. That site was, and still is, the very popular tech commentary site Slashdot. (Joe Trippi posted there in July 2003 to raise awareness about the Dean campaign). This mode of interacting became dominant in the intervening years as personal blogs adopted the form. Harvard’s Berkman Center soon became a big champion of the blogging technology, hosting the early blog conferences, and bringing many of the blog evangelists under its wings as affiliates.

In retrospect, the trip down Drexler’s path was a wide detour. Most of the annotations efforts, save Slashdot, represented the absence of the social. There were too many web pages with no comments; and too few pages overwhelmed by comments, there was never the opportunity for a shared social experience. Microsoft did ultimately introduce annotations as “Web Discussions” to its Office suite. It’s possible it is being used in some closed environments, but it doesn’t seem to be used across open organizations.

What did Drexler have in mind with the “social” designation? His vision of the social, he told Allen for the evolution article, was “the whole of society” that is, “not group-scale, but society-scale.” It is still a noble goal. Though perhaps we are better off being able to consider social software in social terms.

Perhaps what Drexler was looking for was orthogonal to the social interchanges– he was looking to tap into the human initiative for constructing knowledge.

When I started Civilities, the very first post I composed, after the necessary About page, was an illustration of a term constructive media, to set the groundwork for what I wanted to do. I was familiar with the CritLink project, and had developed an annotation tool following in its footsteps; I had probably read papers similar to Drexler’s over the previous decade; I had read the 1995 book The Social Life of Documents, by Xerox researchers John Seely Brown and Paul Duguid many years before. I had pictured that the process should not only applied to academic papers, but to the popular media, to workgroups, campaigns, and organizations. My understanding was that these sorts of groups ultimately needed to focus on producing things: documents, articles, policies, plans, which lead to more tangible things like knowledge, events, behavior, etc.

There is clearly a constructive initiative when Internet users started putting their energy into blogs, which have the permanence that mailing lists do not. There is even more constructive behavior in wiki building, which is by its nature a normative structure as opposed to a narrative one. But they interplay. My analysis of Abuzz was that it was designed as a constructive system that a core group of users ultimately abandoned and used socially. On the other hand, my friend Kaliya Hamlin suggests that some people are finding constructive purposes in the social digital photo sharing service, Flickr. The bookmark sharing services can be used socially (find out who else shares similar opinions) or constructively (find out how all people rate a given resource).

There may be a Yin and Yang balance of these models. Though I suppose that it may shift over time. Without question social software should be used to help disparate people find each other, and to improve the social cohesion of existing organizations. Once the organization expands, and relationships get strengthened, that’s when the group needs to concentrate and engage the constructive gear.

Update, Janury 26th, 2005: I fixed several typos in this piece that were two months old, and added some hyperlinks, and split some long sentences here and there. Re-reading the piece, I realize I still need to write a piece on the pitfalls of social software design. I’ll save it for another day. And more importantly, I’d like to come up with a stronger conclusion on why the split happened. But it did and here we are.