John Gruber of DaringFireball.net explains very well why Apple's upcoming Dashboard mini-application-layer is NOT a rip-off of Konfabulator at DaringFireBall.net: Dashboard vs Konfabulator.
Can anyone who enhances something another developer wrote really expect that the developer of the original product won't add that feature directly at some point? Read John Gruber for reasons why Apple would have been throwing money away to buy Konfabulator.
I'm working on a program that will enhance/work with iTunes to help manage your songs in ways that Smart Playlists don't easily do. And yet, I expect that if it eventually becomes popular, Apple will just add it in to iTunes. All that does is validate my add-on, showing that it should have been in the product in the first place.
If you want to truly be unreplaceable, you have to come up with something that can be patented in some way, and is not derivative. Konfabulator was not a new idea, just a particularly artist-friendly method. As a programmer (of a sort) and not a graphic designer, I found that Konfab's virtual requirement of PhotoShop proficiency led me to realize I'd be better off learning AppleScript Studio (since I knew AppleScript already). Without an IDE, Konfab is just a run-time engine. Dashboard will let anyone be able to run apps a developer creates, not just those who paid $25. Oh, and that developer can use Safari and their existing web-developing knowledge to write Dashboard gadgets.
I feel bad for Arlo Rose, but he doesn't have a legitimate complaint that Apple ripped him off. Rather, he developed in an area that was ripe for evolution of the OS. Especially true once WebCore was released.