While going through the periodic find-new-job-which-sucks-slightly-less-than-previous routine, I keep having crazy thoughts about starting my own business.
I don't want to do yet another webdev shop, competing with 18-year-olds undercutting you from their bedrooms. I have a few ideas that I think are marketable; some of them are completely original, some aren't but their market is potentially so big that it doesn't matter.
One falling in the latter category involves mobile phones, and how to use them in companies to streamline certain processes. To do this kind of things, you don't need much on the phone UI; random sync with some sort of networked server is easy to do.
I've looked into Python for S60, but unfortunately, while potentially fantastic, is not really there yet in terms of adoption and ease of deployment. The only realistic choice is J2ME; to my surprise, it looks easy enough to build what I want with it, and pretty much all recent handsets support it well enough for my needs.
So I installed EclipseME and the Sun Java Wireless Toolkit, and I'm already 40% towards having a working protoype. I just need to wire it to a simple django app at the other end, and then I could really sell it, or at least do a round of funding (there are some fantastic government schemes for small new businesses here in England).
What scares me at the moment is setting the right price for it; it's not the kind of thing that will save massive amounts of money to a company, but it will speed up some operations. I need to quantify these improvements in order to have a strong sales pitch to set the right price, and to be able to build a proper business plan which will take into account development time, expenses, etc. I also need to catch up fast with the main player in the market already (a small Portland-based operation) which has a head-start but, I think, not yet the amount of money required to break it big in the European space. As I said, the market itself looks big enough for more than one player, and after you establish relationships with customers you can go and build new things for them.
In the meantime, I'm using friends as focus groups ;) and the feedback is good. A colleague even offered himself as part-time head for development, which is something I'd really want before doing the jump.
In all this, the folks at GeekUp are being really supportive. I wish I'd found that list 3 years ago.