% fortune -ae paul murphy

Learning from stuff that works

Look at the big trends in IT adoption, the stuff that works for millions of users worldwide, and you see that these things have three things in common: a centralized, shared, server architecture; a focus on facilitating communication between users; and, the ability to reduce or eliminate the threats implicit in heirarchial communications.

Whether it's photo sharing via Flickr, community collaboration via something like the new Kanosis offering, or an ad hoc protest organized via simple text messaging, all of these things centralize processing, push functional control out to the user, and treat the user device as no more than a display and input host - something that's often achieved by creating a custom client whose real job is to stand in for a standard interface while isolating the interaction from other software on the device.

What's most striking about this is what it's not: it's not Microsoft client-server, it's not data processing, and its not unreliable. In fact users trust these tools enough to learn their limitations and then push the boundaries to find new, and often genuinely innovative, uses for them. As a result text messaging usage, for example, now perfectly meets the definition of art: compressing maximum meaning within the tightest possible structure.

The second characteristic these successes have in common is that their function focuses on communications across both time and distance: Dick, in Canberra, can file a message received by Jane, in Ottawa in near real time and retrieved by Spot, in Toledo, a week later.

What makes this kind of computing function so powerful is that it acts as a kind of communal memory, letting people do better in groups than they can by themselves. Thus Dick and Jane may be good at whatever they do, but working together they produce something better -and accomodating Spot's ideas a week later gives them think time, a third party review, more incentives to get it right, and usually an even better long term result.

The third shared characteristic is freedom from heirarchial presures -whether achieved through real anonymity, peer group self selection, or simply by size the social risk associated with effective use of these tools approaches zero.

As individuals humans are generally pretty smart, put them in heirarchially controlled groups and perceived social risks cause them settle at the lowest common level, but put them into eglatarian communities without time limits and personal risks and what you get is Linux, open source, new political movements, and intellectual honesty. In other words, take away the social controls while extending communication over time and what happens is that group think effects become positive, not negative.

I don't know if Kanosis is going to be a market winner or not, but I do know this: collaborative groups can amount to much more than the sum of their parts -and collaborative software based on centralized switching with decentralized control, has been and will continue to be, a big part of facilitating that effect.

And there's a direct bottom line consequence for business here that leads directly to an action recommendation for IT: an anonymous forum set up for use by people in your organization can share all three winning characteristrics.

Put a bunch of middle managers and engineers together in a room to assess some radical new ideas, and what happens to those ideas will be determined by the most change resistant people present; but put those same ideas in front of that same group using an anonymous forum and you'll get a much more honest discussion, useful suggestions for improvement, and ultimately a much higher rate of acceptance.

In other words, it's time to bring this kind of computing in-house -and there are at least a dozen good software choices for anonymous forums, all of them open source, and all of them free.

Paul Murphy wrote and published The Unix Guide to Defenestration. Murphy is a 25-year veteran of the I.T. consulting industry, specializing in Unix and Unix-related management issues. groups are