Do me a quick favour. Take a moment and think of the exact opposite of a traditional Web portal — conjure up something like AOL or, in the non-profit space, something like TechSoup — and then think of the opposite and tell me what comes to mind? If you come up with something concrete, pop your thoughts in the comments below.

I’ve asked myself this question repeatedly over the last couple of years. I think the first time was when talking with Mark about the online strategy for Telecentre.org in 2005; we probably joked around with silly terms that rhymed with “portal” while discussing the practical realities of what it would mean to embrace the spirit of the “anti-portal.” And why bother? Because even the members of the Telecetre.org community had come to the realization that portals aren’t community-building tools:

“telecentre.org should not create ‘a big portal’. Rather, it must find ways to build on and extend existing Web sites, mailing lists and other online resources for telecentres” — quote from the Telecentre.org Online Strategy, taken from a workshop flipchart in Accra

Early this year, after a workshop, I stumbled on the topic again while talking to a participant from a large grant-making organization.

“We’re working on a portal project. What do you think?”

And, I found myself naturally going back to that same answer: What you should really explore is the exact opposite - the anti-portal.

Last week I got to sit down with that organization to have a frank discussion. A discussion about what that might really mean in practice, or what it would look like for an organization that was serious about trying something different… Something the opposite of “big” and, instead, something aggressively small.

The key ideas came back in a flash. To be successful at community-building, knowledge dissemination, network facilitation, and — most importantly — to be sustainable, something that is the opposite of a portal would have to be:

• Edge-focused: Sustainable networks often demonstrate strong connections at the edges. Strength at the edges means that the centre — often the early glue of the network — has receded into the background, and that the edges are able to communicate effectively without central organization. Think of a doughnut vs. a Timbit.

• A working system, not a logging system: One of my first realizations that most software is inherently flawed came from a demonstration of an archaic system that had been built by Ellis Don for managing construction projects. Though the system was a monstrosity, it achieved a mystical stature in my memory for being one of the first “working” systems that I’d ever come across. A working system is one that doesn’t require a separate act to create a log — or an artifact — of the act of working. Most software we use today relies on the logging paradigm: you need to enter the information in as a separate act after you’ve done something. E-mail is probably the most straightforward working system, not only is it simple, but it logs the fact that you’ve used it by putting your message in the Sent folder. This paradigm would need to also be present in a successful anti-portal project.

• A seeker, not a collector: Following the lead of initiatives like Google’s Open Social, successful initiatives in strengthening community online will need to seek out the relevant conversations and build bridges between them. People are already having the important conversations that you want to help convene — find them, document them, and act as a shepherd for newcomers, guiding them to the most relevant threads and topics.

• Prioritizing recency over taxonomy: How many times must we see outdated database carcasses lying around the InterWeb? Please people! Even well-intentioned “Database 2.0” initiatives are finding themselves nothing more than virtual ghost towns (possibly pointing people to old, outdated, information). Quite simply: the more recent the information, the more valuable. Well-organized and categorized information is useful for about as long as it takes to file it, and then it’s already heading toward the information dust-bin. Look at relatively straightforward and self-updating examples like Non-profit Matrix, or newer initiatives like the Social Source Commons for guidance on how to approach directories and databases.

• Offering depth vs. accuracy: Finally, from the wise words of people who live, sleep, and eat “search” — the No. 1 way that people seek out information — be sure to balance accuracy and depth. Offer lateral search paths, and non-traditional ways of exploring large collections of information. Think about information discoverability, but don’t neglect the role of serendipity in helping people stumble on the unexpected and timely.

• Iterative and nimble: Coming back to the aggressively small comment, the anti-portal should be a collection of micro-applications that combine to make a larger whole. Not “portlets,” no. But widgets, or gadgets — little magical Web-service consuming bits — that can be quickly tested with the community and either adopted or disregarded. The technology landscape is changing too quickly for complicated approaches to complex problems. Quick, low cost, iterative sprints are the antidote to large requirements documents and RFP-itis.

• Push and pull: And, no matter how anti that anti-portal is, people will forget about it in no time flat. Think about ways to push out information as much as it’s pulled in. Think about all the different mediums that people use to find, consume, and interact with information — face-to-face events, teleconferences, printed guides, audio, video, and more — and don’t limit yourself to just e-mail and RSS.

I’ve tried to stay 10,000 feet away with this post, so as not to jump to tactics before the strategy is fully baked. It would be easy to say “our non-portal strategy is Facebook!” or “we’ll simply create a shared social bookmarking tag and aggregate the results” — but that would, again, miss the point that no single approach is going to act as the “glue” for a community for more than a short time (perhaps a year, or two at most). So, as I try to prioritize that list above — for myself, and for my future clients — I keep coming back to: be small, be nimble, and adapt like crazy.