The cloud is intriguing. I’ve already lived through the private cloud. 30 years ago when timesharing companies let me run manufacturing and sales remotely. It wasn’t as elegant as it is now, but it worked. And then came lower cost minicomputers and eventually PCs. As we all know, PCs changed everything, and they may well change everything again, but not if we let the cloud-only rhetoric go on without challenge.

The current edition of the Economist (Oct 17th-23rd 2009) ran an article on the coming cloud-computing wars (read it here). This article doesn’t take the longer view of technology. The cloud is not binary. The PC’s time has not passed. Why? Distributed computing. If you look at the power of PCs they have gone up dramatically and turning them into dumb clients for smart clouds is a waste of resources. Any client software that uses cloud services (think searching clip art of servers with Microsoft Word as a very basic example). Along with timesharing, I also worked on early distributed computing applications, including a Hughes Aircraft project on distributed manufacturing where intelligent controllers on factory floors negotiated with each other to replan factory when a machine broke down or a part didn’t show up.

PCs provided individual control over data, but even better, they provided local processing power. That local processing power has now been distributed into phones, media players and cameras – but not in an intelligent way. Those devices are mostly very dumb clients to the PC or the clouds. They are getting smarter, but all of this computing should not just end with devices as end points to cloud agendas. The power of the PC was to let people set their agenda. With distributed computing, individual PCs could act as nodes on an integrated network where problems could be solved faster and more creatively than if they were tackled by the cloud alone. But because development efforts are being split between legacy clients and the cloud, this huge area of distributed computing isn’t really being explored effectively.

Complex modeling of economics, science or physical objects is often accomplished using proprietary distributed computing (think rendering farms for animated movies and video games, weather forecasts). But that collective problem solving hasn’t been democratized to making sense of the information on my computer or your computer. It’s great that we have wonderful animated features coming out of Hollywood.  But I would like to see my computer be part of a network that makes it smarter every time I turn it on. It can’t do that by itself as effectively as it can using all of the processing power available, and the cloud can’t do it without the intimacy of data on my phone and my PC. To take knowledge work to the next level we need to turn our admiration for social computing on to our machines and let them get social. If we can crowd source the next great marketing idea, perhaps our PCs can figure out what they need to know before the next meeting – before we ask for it. I’m not convinced the cloud will get me there alone. I like having something that is mine act as my agent, and I want it to bring its friends along for the insight.