I'd forget the the "writing cool stuff" part as much as possible, and focus as much as possible on "getting things done" and "making some money" (where "making

I'd forget the the "writing cool stuff" part as much as possible, and focus as much as possible on "getting things done" and "making some money" (where "making some money" is short-hand for actually shipping code that somebody is willing to pay for.) It's pretty easy, standing on the outside, to see where you're going wrong. Let's start here: I'm currently writing something SMTP based. First I started using Twisted and Python, which includes its own SMTP implementation, So far, so good... then I went on looking at scaling, Warning, Will Robinson! Warning! Forget scaling at this point-- scaling is a luxury problem, as it means you have enough customers that you actually need to scale. For now, focus on getting a first version out to the customers, and when the rate of uptake is significant enough, then you can start thinking about scaling. In other words: always aim for the minimum. That's why they call it an MVP. Try to figure out what the absolute least amount of code you have to write is, to give the users an experience that they'd be willing to pay for. Then you have something to build on.