Around October 2006, two lines began to converge. It appeared that within a few months they would meet, cross, and continue on divergent paths. Those of us paying attention really wanted this to happen, but alas, it never came to pass. After the turn of the year, the lines began to level off, resolving onto seemingly parallel courses. Now, in mid/late 2007, their rate can only be described as geological.
These lines represent the popularity of IE6 versus IE7, plotted over time on Sun.COM. Had the lines crossed, it would have signified the glorious, long-awaited death of IE6. Of the fifteen or so browsers we try to support, IE6 alone accounts for up to three quarters of our debugging workload, plus it makes us write inelegant code. It's not that IE7 is so great, it's just that IE6 is such a hornet's nest of nasty bugs.
Interestingly, every weekend, IE6 usage drops, while IE7 spikes. It gives the line a squiggly appearance. My theory is that IT departments are refusing to upgrade to IE7 for fear of breaking horribly-coded internal legacy web applications. In fact, my wife even confirmed this very state of affairs at her own company. People are locked in. One shudders to imagine the insoluable mass of nasty code grinding away out there on the world's intranets.
I think the desire to upgrade exists, but the hard reality of legacy web apps is keeping things at a standstill. Plus, at the end of the day (literally) most people are comfortable using the same browser at home as they do at work. If you peer at the graph, you can almost imagine the spikes as being little hands grasping out between the iron bars of corporate IE6 imprisonment, moaning, "Save us!"