Just do it
Eric Sink has some terrific advice:
You could spend a lot of time trying to figure out if this idea is any good, and at the end, you wouldn't really know. Alternatively, you could spend the same amount of time implementing this idea, after which you will have the opportunity to really find out if it's any good.
I could link this over to my own struggle to get my idea out of the garage and (at the very least) sitting in the driveway where my neighbors can think up polite ways to ask "WTF?"
But really, this principle is so simple we should look for it in far more places. Isn't this the foundation of iterative development? Can you make a wireframe or mockup or prototype faster than you can draw a pretty UML diagram?
What about selling agile practices? Which ones can you just do
? Wouldn't it be easy to say "we've been trying X
for four weeks and we've achieved Y
"? Isn't it far better to risk asking forgiveness than campaign for permision?
Why not just try pair programming? What's stopping you from just doing test-driven development on your own personal tasks? Is there any reason you can't use Cruise Control
on your workstation even if the team doesn't buy into continuous integration? What's wrong with writing all of your tests in FITnesse
and sharing the results?
I'll make you a deal. I'll stop wasting time writing this blog entry now
if you will go ahead and think of one thing
you could just go ahead and do.
You've got your idea? Ok, here's my 'end' of the bargain.