48 HOURS???!!!

[swestrup]

I must be insane. I've gone and signed up for Blitzweekend. Its an unconference where folks try to roll out something innovative within a 48 hour period. Wireless access and food will be supplied, and there will be keynote addresses and the like.

On the one hand, I'm really excited by the idea, as I haven't really participated in any after-hours hacking sessions (in the original meaning of the term) in ages and from that POV, it should be lots of fun. On the other hand, I usually think of myself as a Professional programmer, not a hack. To quote M*A*S*H's Charles Emerson Winchester III: "I do one thing, I do it very well, and then I move on."

So, this will be a change of pace for me. I don't imagine for a moment that I can create a viable startup in 48 hours. I've been in too many startups and have too much experience with the sheer amount of work involved to imagine I could launch one in that time. I do, however, think its possible to create a prototype of some new core technology that a startup could be built around, and that's what I'm going to be looking to do. The big problem is to identify some technology that A) one could build a startup around and B) is simple enough to do a prototype of in 48 hours.

I have already thought of 3 different possibilities:

1. Zillow done right. Actually, having looked at a bunch of attempts to replace the broken MLS System, Zillow isn't far wrong. But it could really be improved on, and there's a need for an international version. It would involve hacking together Google maps, Craigslist, an MLS screen scraper, various statistical databases, and a wiki engine. On the pro side of things, there's an obvious business model, and an existing market you can point out to investors. The major con for this as a startup is that it would be a target of frivolous lawsuits (as is Zillow) because Real Estate associations would feel threatened by it. And it would need a really good web interface. I hate doing those.
   
2. Memedaq. This would be an online game that would cost pennies to play, but could conceivably rake in big bucks if I could a) make it addictive enough and b) convince major advertisers to play. It would definitely need a web site, but that could be fairly simple at first. The pain would be in wiring in Paypal and populating it with enough entries to make it look worthwhile. Oh, and I'd probably need to hack in a copy of khtml2png or something similar. On the pro side, I already own the Memedaq.com domain and have been meaning to write this for ages. On the con side, I'm not sure if this would be very impressive as a product to show off until it had hundreds of users. At least the web site could start off very simple.
   
3. PSS, a replacement for RSS that uses push via XMPP technology. This is where I am currently leaning. RSS simply doens't scale, and there would be lots of demand for a competent replacement that did. There are a HUGE number of applications and adjuncts to this that one could put together, and I've long wanted to do this. I would need to write (or modify existing copies of) XMPP and RSS servers and clients. I could have a bare bones website for registration (ie, a single page), and it could be done very quickly (I hope). The idea would be to get a preliminary version working as early as possible, and then see how many of the features I can tack on in the time remaining. The pros are that I know of some specific technology that I could use to do this that most other folks will never have heard of, so I have unique things to contribute. The major con is that I'll have to be very careful to pare the concept down to the simplest core that I can write in an 8 hour day, and still be able to add to. If I screw that up, it'll never be finished in time.

Of course if someone else wants me on their team for something, I may not do any of these.