Rapid Application Development(RAD) is all the hype lately, with claims of being able to churn out functional applications in under two weeks. One popular technique is to lock end-users and developers in a conference room for 10 days and build the application on the fly. Another technique sees a cycle of iterations, adding features, showing users, gathering more requirements until you have a workable application. But, sometimes the end result doesn't look anything like what the end-users actually need.
I am asking, what ever happened to go old requirements analysis? I know that doing requirements properly takes time and some organisations view it as barrier to developing applications quickly and in a cost-efficient manner.
Part of the problem is actually who you might have gathering the requirements. Often business analysts are pushed into the role without a real understanding of the tools, technologies or platforms.
So I say if you want good requirements, let the development team gather them. What they lack in analysis skills they more than make up for in knowledge of what is and isn't possible given the time, resource and platform constraints. You may just be suprised at the results— an application that meets end-users requirements that can be developed quickly using your chosen platform and tool set.