Eric Sink wrote great article Exploring Micro-ISVs about running one-person company. I am trying to run very similar ISV now, and some advises are extremely useful. Especially this one:
Don't start too big
In any software company, it's important to find a way to keep your 1.0 cycle as short as possible while still building a product which will generate revenue. This is a delicate balancing act, I admit. If your 1.0 release is light on features, fewer people will buy it. If you build the product that will appeal to the bulk of your market, it will take too long. Where's the happy medium? ... The purpose of 1.0 is to help pay for the development of 2.0, and so on.
I am going to cut off some features from the first release after this article. Thank you, Eric!