This looks cool but I'll be the guy to tell you that doing this is going to be really hard as a project, and as a company. To decently organize different subje

This looks cool but I'll be the guy to tell you that doing this is going to be really hard as a project, and as a company. To decently organize different subjects you won't be able to stay at the course level for long - too large a grain size. Maybe you can stay at university subject and undergrad/grad distinction, but you'll quickly find edge cases may outweigh standard cases when you define further. You're going to have to start parsing the syllabi of individual courses, which means you're going to have a crash-course on the subject. It'll take time, and you won't get much faster at it in another subject. This is assuming a practical approach rather than a more research driven approach, which would take a lot longer. That's the first challenge. The second challenge is that tons of colleges are entering the fray and a LOT of courses are going to get put up soon. So the 700 courses becomes 1500 becomes 5000. You may think this is an opportunity but its at least as much a burden. And once you classify them, they'll change, and you'll have to update them. And you won't have access to critical feedback loops on why they changed or their quality. Between those two challenges you're going to have to find a sustainability model fast. As you note, many others have started to do this, and although they didn't do the particular feature you're talking about w/ matching courses, there's probably a lot to learn from them. I think everyone was going 'platform!' but that involves a LOT of building audience before getting any possible leadgen revenue. At a much higher level, college itself is going through a pretty big rethink. Its worth asking yourself the question: "why isn't there a standard way that colleges organize their courses today?" (hint: its structural, not just pre-internet). Its also worth thinking about how MOOCs are adjusting their offerings (e.g. shorter lengths, highly multidisciplinary content, topical content, etc), and whether or not standardizing across traditional college schemes is today's and tomorrow's problem, not just yesterday's. I think its worth learning a bit about library sciences (the nature of classification), talking to MOOC instructors about their experiences, talking to other course listing startups, & talking to colleges about how they think about what makes a college education (or at least read what they have to say). This probably won't require code (gasp, I say this on HN), it's more a get-out-of-the-building problem. Code is kinda your enemy right now, it'll trick you into thinking you're making progress. The desire to tackle organizing it is noble, but the scope is big. There's lots of other edtech problems out there if you're interested in the space. And all that said, feel free to prove me wrong! Source: a long time in edtech, a lot of time spent classifying (e.g. edsurge.com/products), and been involved in lots of discussions with open courseware folks about these very problems.