Only implementing the features you really need and allowing new ideas to churn around in your head for a while before you implement them is a good way of ensuring that you end up with a simple, elegant product. I believe that constant re-prioritizing is as necessary and quite normal.
This means we shift from building Tank to building sites with Tank. I read somewhere else that it is important to eat one’s own dog food. It helps you see things from a user’s perspective. We’d like to expose people to the system to challenge our framework. This is not design by committee - this will help test our ideas and assumptions to see if they stack up.
and where exactly did you read that it's important to 'eat your own dogfood' then? one of those hippie sites? :)
Actually I think it was somewhere in the 37-signals-sphere. Might have been an interview with the campaign monitor guys. Not sure ;)