I've been reading "The Future of Management" by Gary Hamel lately. I have a staff of nine developers and, while I believe I'm a pretty good manager, I also believe there is always room for improvement. So I try to keep my eyes and ears open for ways to improve, and reading this book among others is one attempt to do so.

One particular concept I found intriguing did involve evolution. He referenced the work of Stuart Kauffman in his book "At Home in the Universe: The Search For Laws of Self-Organization and Complexity." Kauffman describes a "fitness landscape" that includes a mountain range with a number of peaks, each representing higher levels of evolution. A species starts out in a deep valley, but as it wanders it moves up one of the peaks. As it does so, however, the amount of land above it diminishes while the amount of land below it increases. That is, as the species moves up, the chances that the next change in direction will take it lower rather than higher increases. In fact, once a species reaches the summit of the particular peak it's on, it's not possible for it to begin to climb the next higher peak until it descends a bit from the one it's on.

For whatever reason, what that instantly reminded me of is software development tools (where I believe a form of evolution does occur). First there were mainframes, and when they hit their peak, suddenly tools like PowerBuilder showed up and carried us over to the next higher mountain. There was a time when using the new development methodology required re-architecting solutions for problems that the mainframe guys had solved long ago. So we essentially have to start off at a lower point than they were already at. But eventually the tools matured, and we ended up at the top of the next higher peak.

Then along came the web, and other tools carried some developers up another peak. Whether it was a higher one or not is perhaps a matter of debate. I envision it as simply a different peak, no higher or lower than the client/server peak. The problem came in when tools like PowerBuilder also attempted to take people up that other peak, while maintaining their position on the current one. For awhile, that involved quite a step back down the mountain. Lately I think we're finally getting close, but haven't reached the summit yet. It's still not as easy to develop a web-based app using PowerBuilder as it is a client/server app, but it's sure a lot easier than it used to be.

Meanwhile, though, I believe there's another peak looming on the horizon, and this one does look to me like it might be a higher peak. I'm referring to rich Internet applications (RIA), applications that are in a sense a hybrid of client/server and web applications. They attempt to provide the advantages of both platforms without the disadvantages of either. Adobe has Flex, and via AIR can run on the desktop. Google has the GWT and more recently Gears. Yahoo has their YUI and just introduced BrowserPlus. Mozilla has Prism to allow web applications to run on the desktop and Microsoft is trying to get into the game with Silverlight. The old man of the group is most likely Curl, which has been around since 1995 and became a commercial operation in 1998.

The other issue that many of the HTML / JavaScript / XML approaches suffer from (as well as the current PowerBuilder implementation of Web Forms) is browser compatibility. The whole reason that services such as Browser Photo and Browser CAM exist is because it's difficult to write web applications that work correctly for the wide variety of operating system and browser choices out there, and perhaps even more difficult to test the app on all the various possible combinations to ensure it operates correctly. What's particularly nice about the plug-in based approach used by Flex and others (e.g., Silverlight) is that the output is rendered identically (or should be) regardless of the browser or operating system in use.

Most likely PowerBuilder's attempt to scale this new peak will be through Microsoft's Silverlight. However, Silverlight has just started on this climb, and so we'll be looking at another walk down the mountain for a while before we start hitting the heights again. All the same, I'm looking forward to that hike.