The issue of support is a big dog. You need to be able to sell a lot more copies than you will have users who bug you. Do you already have a company framework set up (even if for another purpose) which can be used to field phone calls?

Just as importantly, is it a project that sells itself? Is its purpose so crystal-clear that everybody just has to rush out and buy it? If not, you need to be sure to include enough money for shmoozing magazine editors and creating marketing folders.

If you can build it in two months, why can't someone else? Time and again, open source developers duplicate payware products and take away many paying customers. Not all of them, because there will always be droids who will not touch open source and/or work only in Doze.

If all you want to do is to support yourself, the suggestion to release your core code, especially with a gee-whiz save-the-world application that gets you publicity/notoriety, is a great one. You can then get work customizing your framework for paying customers or writing books about it, like the R* on Rails guys.

Be careful. Betting your future on business is just as perilous as betting the stock market will go up when you need it to. I've burned $300K+ on one business, and I'm working hard to make it back (and more) with another one. Be sure you're ready to take on all the other attributes of a successful business (money and time management, taxes, promotion, marketing, etc., etc., etc.) before you go there, and be very sure you have a product which is as good as you think.