There are only a few web apps I think can be useful within this 3 hour construct. I think it's much better to develop a web app that takes about a weekend or maybe 2 to complete. You get something with a lot more value and can truly extend your ideas outward. Furthermore, a weekend still provides the construct of "fast."<p>Building a web app in 3 hours will leave you at best with something like threewords.me or its clones. While those can go viral, they rely heavily on luck. Why not spend 2 hours and build a beautiful and captivating launch page for an idea that would take you a weekend or two to build. Then, spend 1 hour reposting it everywhere and try to get feedback. This way, you can see if the concept itself gets any traction.
While I agree that it's probably impossible to produce a whole web application in 3 hours, I think the point here is more about incremental delivery_. Note that in the post, OP is actually talking about 3-hour <i>features</i>, not entire apps (after step ‘show to people’ you can go back and develop another 3-hour feature, if you like). And I doubt there exists a web application (or almost any piece of software, really) that couldn't be developed this way.<p>.. _: <a href="http://c2.com/cgi/wiki?IncrementalDelivery" rel="nofollow">http://c2.com/cgi/wiki?IncrementalDelivery</a><p>Problems, however, arise when your idea requires specific domain knowledge which you haven't got yet; it's likely that you'll need some time to become comfortable in this particular area. Though I doubt it should necessarily be counted as time spent on app development.<p>And there's always some in-brain idea development. Say, you have an idea of blogging via email. You cook this idea inside for a few days, working out some details in background, then sit down, code for a three hours, and boom, you have posterous. Then you see that people don't care about email blogging, but they care about bells and whistles, so you sit another 3 hours and create basic skinning support. It's better than when after two weekends on email-blogging you discover that you were going in wrong direction. (Of course, I know nothing about real Posterous development, I just tried to clarify.)<p>Also, I think it's generally useful to be able to produce something in 3 hours. You need to keep your tools hot & ready.
This is a good exercise to help train yourself to strip out the less important ideas/features. Do it enough and apply the new learnings toward a bigger project.