Sometimes when I start a small project (like an android app), I don´t know which approach will work out at the end, and I just go for one approach and give it a try. But if I never used this approach before (for a sort of application I´ve never programmed before) it is like stepping into unknown terrain. I don´t know which libraries to use (maybe I have to try out several libraries) and there are so many unkonwns (like: how to get raw audio data in android)

So then my development process goes like this:

• Write a piece of code to see if the approach has a chance. (The more uncertain the approach is, the uglier the code gets)
• If it works, refactor a lot until it is beautiful

I think it could be a waste of time if I planned my software design in detail at this point, it would be like planning a trip without a map.

Is this part of aglie development? How do you deal with unknown terrain in software development?