Should be simply a port of existing software. I imagine a community of people sharing algorithms about many different subjects and at different levels of seriousness. My brother wrote a paper on Math Equations for the Socially Challenged in college that would fit well into this concept.
This market is probably sewn up -- an original algorithm can be the core of a thesis, while a useful one should be the kernel of a business plan. So you're competing with arXiv.org and YCombinator. Good luck?
I am actually talking about providing the tools to a community of users that will create those things to share with one another. arXiv.org and YCombinator address completely separate needs.