Academic hackathons should have a different focus from VC-based ones. With VCs, the goal is to get your project funded, or to become more visible to VCs so that when they form new companies they might recommend you to the founders. A hackathon might be seen as a job interview, in a way.  

An academic hackathon, like the ones held at Yahoo a number of years ago (where the concept originated), might be less like a job interview and more like a break from the usual work routine, to shake out new ideas or approaches. People who work or study in the same institution get together for an intense period to share. A chance to do collaborative experimentation. To try an idea out and see where it goes. See if you can't solve a problem that's part of someone else's work.  

In 24 hours I wouldn't try for a breakthrough. But it is possible to focus for a few hours on a problem and come up with alternate approaches. Or to rework something that fell by the wayside in the past because of scaling issues perhaps. Or some piece of the puzzle was missing last time it was attempted, where a solution now exists. 

Anyway, with that in mind, here are few ideas that you might try if you find yourself at an academic hackathon, such as the one this weekend at NYU. If anyone makes progress on these projects I will write them up on Scripting News, of course.  

1. A few years ago I had a site called Share Your OPML. I had to take it offline because of scaling issues. The idea is that members of a community submit their OPML subscription lists to a central server. It mashes them together into a database and provides a number of readouts. It tells you what the most popular feeds are of course. But more important, it recommends feeds based on what other people are subscribed to. I think this project really could go somewhere. But this is where it starts. You can find SYO in archive.org's wayback machine. 

2. How to display a River of News? Here's a bare-bones river. It's a relatively realtime view of the NY Times RSS feeds. For each item there's more metadata, but not much more. (The usual elements of an RSS item.) People say this display is plain and unexciting. They're right! But it looks good on a Blackberry. (I use an iPhone nowadays.) So it wouldn't bother me if it looked better and did a little more. But not a lot more. The simplicity of the River is central to its utility. Your mission, if you decide to accept it, is to prototype an HTML display of this river, perhaps with some prefs that a user can set, that is more beautiful. You might be able to do that in 12 hours or so. If you have an idea, maybe even less.  

3. A project related to #2, write a client for the FeedHose protocol that I released last week. You can get the NY Times flow that way, without having to read the RSS feeds. That way your mockup will really work.  

4. A really ambitious project that you could start at a hackathon is a centralized subscription manager. It would have a bookmarklet that installs in the browser, and would connect back to a server that manages subscription lists. Think of it as XMarks for feeds. Even the XMarks guys didn't realize how hot their own product is. This one is even hotter, if it can get support from people who do aggregators. Even if that never happens, it's still an interesting project to know how to do, just in case.  

5. Hack status.net so users can edit the templates, producing a Twitter-like service that integrates with the look and feel of your site or blog. I would love to have one here on scripting.com, but not if it looks just like the status.net they operate in Montreal. I want mine to look like mine. Take the design features of Tumblr and apply them to Twitter. There's a big creative group that's been left out of the microblogging world, designers. Bring them in, lead them. For prior art look at how Tumblr does templates. 

6. Readability is open source. Apple did some great stuff with it, as did (little-known fact) Flipboard. Instapaper is just getting going as a business. But these products get by by scraping. What if you created a high-level content platform that a blogging tool maker like myself could generate content for. What higher-level constructs would you implement? Could we get closer to Ted Nelson's vision for hypertext? (For extra credit, who is Ted Nelson and what is his vision of hypertext.) 

7. I want to hack DNS. Implement a simple efficient REST interface for a domain name server. Let me add and delete records. Fully manage a domain via API. The API for DNS is ancient. Modernize it. (If your really focus and if you know how to program DNS, this could be a 12-hour project, imho.) 

That's about it for now. If one or more of these ideas is done at the NYU Hackathon tomorrow, I'd be very happy! Let me know.