I just screwed up. I accidentally marked all of my Google reader items as read when what I intended to do was mark a specific folder as read. If this was a desktop application I could hit undo and I'd back up, have all of my data the way I want it, and be ready to read all of the delicious feeds I've been missing this week. But Google Reader is a web application, so I can't do that, I'm just SOL.
This is one of the huge, huge problems with web applications that we haven't talked about much. Web apps are great. They're easy to get to, they store your data in the cloud and they're fairly ubiquitous. But we haven't gotten to the point where we can replicate true desktop paradigms on the web. Because the web was primarily a document medium, we didn't have to think about application problems like states or offline issues. But now we do and while we're making a lot of progress, there are things keeping us from fully making the jump.
RIAs haven't solved this problem either. But because RIAs started with the idea of building real applications on the web, we're closer than other technologies. But it's funny because the forces that pull us in the direction of desktop applications also pull us towards the web. If you figure out a very robust way to incorporate states (so you can add undo functionality) then how do you deal with people that want to deep link into an RIA? Should you be able to link into each stage of the undo stack?
Technologically, blending the web and the desktop isn't hard. It's blending the features from each that's difficult and it's why RIAs are such an interesting problem. We're all still figuring out what to take from the web world and what to take from the desktop world. I want Undo in my rich internet applications. The question I what I need to give up to get that.
Image courtesy of NewbieGamerBlog.