This is wonderful. I'm totally gonna use it and contribute to it on a long flight tomorrow.
You can always tell a good idea when it gets your gears turning. With the sort of database that you'll be building, you could experiment with different visualization strategies and hyperlinking between cheat sheets!
A simpler task for version 1.0.1: Make sure that formatting is cleaner on the generated PDFs. For example, on this cheat sheet there is one hanging row on an otherwise empty pp2
Pick something you want to make and make it. The best guideline for this task is to scope small. For example, try and code a clone of Breakout. Now try and do it in a language you've never used. Maybe try and render it server-side and send the info to a canvas element on a web page.
The other direction you could go is implementing algorithms in various languages as a sort of mental workout, but that's a lot less fun.
I'm going to assume "Other than x has a bigger market" can be extrapolated to be "Other than x has more of an upside."
And from that I'll extrapolate that you are in the entrepreneur game to learn and grow as an individual as well as provide something new and interesting for its own sake.
So I'd say "Pick the one that would likely cascade to greater discoveries down the road." In this case, a new mathematics would be a better choice than a new engine based on understood mathematics.
I don't know if this advice is any better than "just Google it" but I usually come up with every single variation on the idea that I can and just start brute forcing search terms. By the end I'll have a computer-slowing amount of open tabs. I bookmark them all and then start going through the results. It's rare that I don't find some usable code.
I'm a video game producer that has published on basically every modern platform. The best advice I can give is to laser-focus on your core mechanic, always asking yourself "Does this add to the fun?" Everything else follows from that.
Different platforms have different customers, demand different games, and have different challenges. For a small project, I highly suggest using a sophisticated and cheap multiplatform middleware engine like Unity3D. It allows you to quickly and easily iterate without worrying about coding new technology.
If you are looking for funding on a small project, I suggest something like Kickstarter. Heads up, unless your project has a sophisticated online component the biggest money sink is going to be your art. If you can a) get users to generate the assets, b) generate the assets programmatically, or c) use an aesthetic that revels in its own lo-fi chic you will have cut your costs dramatically.
The biggest hurdle will be your own enthusiasm. For a small project, keep the scope small. Check out flOw and World Of Goo to get an idea of what I mean by "small."
The best way to find something that is both new and fun is to experiment. Come up with wacky ideas for mechanics and build really ugly prototypes. When you have an ugly piece of unstable software that manages to enthrall its players despite itself THAT'S when you move forward.
You can always tell a good idea when it gets your gears turning. With the sort of database that you'll be building, you could experiment with different visualization strategies and hyperlinking between cheat sheets!
A simpler task for version 1.0.1: Make sure that formatting is cleaner on the generated PDFs. For example, on this cheat sheet there is one hanging row on an otherwise empty pp2
http://www.cheatography.com/davechild/cheat-sheets/javascrip...